Visit https://ebookfinal.

com to download the full version and

explore more ebooks

Probabilistic Models for Computer Networks Tools

and solved problems 2nd Edition Jose Alberto
Hernandez

_____ Click the link below to download _____

https://ebookfinal.com/download/probabilistic-models-
for-computer-networks-tools-and-solved-problems-2nd-
edition-jose-alberto-hernandez/

Explore and download more ebooks at ebookfinal.com

Here are some suggested products you might be interested in.
Click the link to download

Probabilistic models for dynamical systems Second Edition

Benaroya

https://ebookfinal.com/download/probabilistic-models-for-dynamical-
systems-second-edition-benaroya/

SAP R 3 Handbook Third Edition Jose Antonio Hernandez

https://ebookfinal.com/download/sap-r-3-handbook-third-edition-jose-
antonio-hernandez/

Geometric Programming for Computer Aided Design 1. Edition

Alberto Paoluzzi

https://ebookfinal.com/download/geometric-programming-for-computer-
aided-design-1-edition-alberto-paoluzzi/

Expert Cube Development with SSAS Multidimensional Models

Expert Tips and Tricks for Designing Analysis Services
Multidimensional Models 2nd Edition Alberto Ferrari
https://ebookfinal.com/download/expert-cube-development-with-ssas-
multidimensional-models-expert-tips-and-tricks-for-designing-analysis-
services-multidimensional-models-2nd-edition-alberto-ferrari/
More Baking Problems Solved Stanley P Cauvain

https://ebookfinal.com/download/more-baking-problems-solved-stanley-p-
cauvain/

Concepts Models and Tools for Information Fusion Eloi

Bosse

https://ebookfinal.com/download/concepts-models-and-tools-for-
information-fusion-eloi-bosse/

Essential Tools for Operations Management Tools Models and

Approaches for Managers and Consultants 1st Edition Simon
Burtonshaw-Gunn
https://ebookfinal.com/download/essential-tools-for-operations-
management-tools-models-and-approaches-for-managers-and-
consultants-1st-edition-simon-burtonshaw-gunn/

Computer Networks 5th edition Edition Tanenbaum

https://ebookfinal.com/download/computer-networks-5th-edition-edition-
tanenbaum/

Solved Practical Problems in Fluid Mechanics 1st Edition

Carl J. Schaschke (Author)

https://ebookfinal.com/download/solved-practical-problems-in-fluid-
mechanics-1st-edition-carl-j-schaschke-author/
Probabilistic Models for Computer Networks Tools and
solved problems 2nd Edition Jose Alberto Hernandez
Digital Instant Download
Author(s): Jose Alberto Hernandez, Pablo Serrano
ISBN(s): 9781291546873, 1291546871
Edition: 2
File Details: PDF, 2.53 MB
Year: 2015
Language: english
Probabilistic models for computer networks:
Tools and solved problems

José Alberto Hernández and Pablo Serrano

Universidad Carlos III de Madrid
Copyright © 2013
ISBN: 978-1-291-54687-3
 iii

This book is dedicated to all women that

make men better

 Contents

1 Probability review 1

2 The exponential random variable 21

3 The Poisson process 37

4 Discrete-Time Markov Chains 55

5 Continuous-Time Markov Chains 79

6 Classical queueing theory 103

7 Open queueing networks 143

Index 167
Preface

About the authors

José Alberto Hernández and Pablo Serrano are full-time

lecturers and researchers at the Department of Telematic
Engineering of Universidad Carlos III de Madrid, Spain.
In 2009, they designed a course on performance evaluation
of networks for the second year of the Bachelor’s Degree
in Telematic Engineering at the university. This book com-
prises about four years of experience teaching this course.
The feedback collected from students has been very useful
in understanding which concepts require further explana-
tion and examples for a better understanding.
José Alberto Hernández completed the Bachelor’s de-
gree in Telecommunication Engineering at Universidad
Carlos III de Madrid (Spain) in 2002, and the Ph.D. degree
Figure 1: J. A. Hernández
in Computer Science at Loughborough University (Leics,
United Kingdom) in 2005. From 2005 to 2009, he was a
postdoctoral researcher and teaching assistant at Univer-
sidad Autónoma de Madrid, where he participated in a
number of both national and european research projects
concerning the modeling and performance evaluation of
communication networks. In 2009, he moved to Universi-
dad Carlos III de Madrid, where he became an Associate Figure 2: P. Serrano

Professor. His research interests include the areas at which

mathematical modeling and computer networks overlap.
He has published more than 80 scientific articles in both
viii

peer-reviewed journals and conference in-proceedings, in

the areas of network performance and reliability, broad-
band networks, high-speed switched networks, optical
WDM, converged optical-wireless access networks and
energy efficiency.
Pablo Serrano obtained his Telecommunication Engi-
neering degree and his PhD from Universidad Carlos III
de Madrid (UC3M) in 2002 and 2006, respectively. He has
been with the Telematics Department of UC3M since 2002,
where he currently holds the position of Associate Pro-
fessor. In 2007 he was a Visiting Researcher at the Com-
puter Network Research Group at Univ. of Massachusetts
Amherst partially supported by the Spanish Ministry of
Education under a José Castillejo grant. He has published
over 50 scientific papers in peer-reviewed international
journal and conferences. He serves on the Editorial Board
of IEEE Communications Letters, and has served on the
TPC of a number of conferences and workshops including
IEEE Globecom and IEEE INFOCOM. His current work fo-
cuses on the performance evaluation of wireless networks.

Book structure

This book is split into seven chapters, not isolated but de-
pendent, as shown in Fig. . We suggest the reader to follow
the order presented in the figure.
The first chapter provides a basic review of some im-
portant probabilistic concepts and theorems, including
expectation, joint and conditional probability, the Bayes’
theorem, etc. Some well-known Probability Density Func-
tions are also introduced, most of them will be thoroughly
used across the book.
The second and third chapter focus on the exponen-
tial distribution and the Poisson process respectively. The
Milestone 1: Memoryless memoryless property of the exponential distribution comprises
property of the exponen-
tial distribution
 ix

Figure 3: Book contents

and relationships be- 1. Probability
Review
tween chapters.

2. Exponential 3. Poisson
4. DTMCs
Distribution Process

5. CTMCs

6. Classical 7. Open
Queueing Queueing
Theory networks

the first important milestone of the book, since it will play

a key role in the Poisson process, and continuous-time
Markov chains. We often say in class that the exponential
distribution and the Poisson prosess are like husband and
wife, since they go everywhere together.
Then, the book introduces discrete-time Markov chains
which are very useful in the analysis of a large number of
communication systems, protocols and network-related
situations. The Chapman-Kolmogorov equations and the con-
cept of limiting-state probabilities comprise the second big Milestone 2: Chapman-
milestone of this book. Kolmogorov equations
and limiting-state prob-
The next step is to present continuous-time Markov chains
abilities of discrete-time
which combine discrete-time Markov chains with expo- Markov chains
nential distributions. Before addressing this chapter, the
properties of the exponential distribution and the analy-
sis of discrete-time Markov chains must be very clear to
the reader. The third milestone in the book comprises the Milestone 3: state-
ability to correctly use a continuous-time Markov chain to transition-rate diagram,
infinitesimal generator
model a given problem, including the state-transition-rate
and balance equations of
diagram, the infinitesimal generator and the balance equations continuous-time Markov
chains
x

to obtain the steady-state probability vector of a Markov

chain.
Once continuous-time Markov chains are well under-
stood, then queueing theory comprises just an application of
them. Indeed, queueing theory relies on both continuous-
time Markov chains and the Little’s theorem, which is the
Milestone 4: Little’s fourth important milestone in the book.
theorem Finally, open queueing networks builds upon queueing
theory and provide tools to analyse big complex networks
of queues in an easy manner. The fifth and final milestone
Milestone 5: Burke’s and comprises both the Burke’s and Jackson’s theorem which allow
Jackson’s theorem for easining the analysis of open queueing networks.

Focus and readership

The book aims at providing a step-by-step learning, start-

ing from intuitive examples, introducing the theory and
strengthening the learning process with new exercises.
The goal is to involve the reader through the learning pro-
cess with small frequent challenges in every chapter, on
attempts to make him/her feel confident with the theory.
The length of the book has been deliverately intended
to remain short, giving priority to examples and problems
over long theorem proofs. The reader is recommended to
visit other classical books (those pointed out in the bibli-
ography) for further details concerning theorems and their
proofs, since we have more focused on intuition rather than
rigour. The book includes over 100 solved examples and
exercises that cover the main concepts introduced in theory,
with an application to computer networks and communica-
tion protocols.
The contents of this book has been designed to span a
whole semester in the third or forth year of a degree in
Computer Science or Electrical Engineering. In addition,
this book might be suitable for students conducting their
 xi

research in fields related with the performance evaluation

of computer networks.
In light of this, the book assumes that the reader is al-
ready familiar with probability theory and communication
networks, especially concerning layer-2 protocols including
Medium Access Control (MAC) protocols and TCP/IP fun-
damentals. In addition, it is recommended that the author
has some programming skills in Octave or Matlab.

Acknowledgements

The writers would like to acknowledge the help of Dr.

Ignacio Soto, Dr. Isaías Martínez and Dr. Andrés García for
valuable comments and proof-reading.
1
Probability review

1.1 Random variables

Consider a random experiment having sample space S. In

probability theory, the sample space S of an experiment or
random trial is the set of all possible outcomes.

Random variable A random variable X is a variable whose

value is not known. Its possible values usually represent
the possible outcomes of an experiment.

For example:

1. Tossing a coin may have two possible outcomes: SX =

{ Heads, Tails}.

2. Rolling a die may have six possible outcomes: SX =

{1, 2, 3, 4, 5, 6}.

3. The height of adult human beings: SX 2 (150, 235) cm.

Random variables may be continuously or discretely val-

ued. The first two cases are examples of discrete random
variables, whereas the last one represents a continuous
random variable.
2 probabilistic models for computer networks: tools and solved problems

1.2 Probability Density Functions

The Probability Density Function (PDF) describes the rela-

tive likelihood of every possible outcome of an experiment
characterised by some random variable.
In the previous examples, the random variables can be
characterised by the following PDFs:
1. Tossing a coin: P( X = Heads) = P( X = Tails) = 0.5

2. Rolling a die: P( X = i ) = 16 , with i = 1, . . . , 6 (uniformly

0.4

0.35

distributed).
0.3

0.25

0.2

0.15 3. Human height: f X ( x ) ⇠ N (µ, s)

0.1

0.05 It is worth remarking that:

0

 P( X = x ) = 1
−3 −2 −1 0 1 2 3

Figure 1.1: Normal PDF, (1.1)

x
N(0,1).
for discrete random variables, and:
Z
f X ( x )dx = 1 (1.2)
x
for continuous random variables.
The Cummulative Distribution Function (CDF) gives the
probability that a random variable takes a value below or
equal to x, i.e.:
1
Z x
0.9
FX ( x ) = P( X  x ) = f X (t )dt cont. r.v. (1.3)
0.8
•

 P( X = k)
0.7

0.6 FX ( x ) = P( X  x ) = disc. r.v. (1.4)

k x
0.5

0.4

0.3

0.2
Finally, the Complementary Cummulative Distribution
0.1

0
−3 −2 −1 0 1 2 3
Function (CCDF) or Survival function gives the probability
Figure 1.2: Normal CDF. of taking a value greater than x:
SX ( x ) = P( X > x ) = 1 FX ( x ) (1.5)
In other words:
1
Z •
0.9

0.8 SX ( x ) = f X (t )dt cont. r.v. (1.6)

0.7 x
SX ( x ) = Â P( X = k) disc. r.v. (1.7)
0.6

0.5

0.4

0.3
k> x
0.2

0.1

0
−3 −2 −1 0 1 2 3

Figure 1.3: Normal

CCDF.
 probability review 3

1.3 Expectation and variance

The expectation and variance provide a good summary of

the properties of a random variable.

Expectation The expected value of a random variable is

the weighted average of all possible outcomes for the
experiment it characterises. Mathematically:
E( X ) = Â xP(X = x) (1.8)
x

for discrete random variables, and:

Z
E( X ) = x f X ( x )dx (1.9)
x
for continuous random variables.
It can be shown that the
expectation is a linear
Variance The variance is a measure of how spread the pos- operator:
sible outcomes of an experiment are with respect to its E( aX1 + bX2 + c) =
expectation.
= aE( X1 ) + bE( X2 ) + c
Var ( X ) = Â[ x E( X )]2 P( X = x ) (1.10) for any two random
x
variables (independent or
for discrete random varaibles, and: not) X and Y and any real
Z numbers a, b and c.
Var ( X ) = [x E( X )]2 f X ( x )dx (1.11)
x
for continuous random variables.
The variance is not a
Alternatively, the variance may be computed as: linear operator:

Var ( X ) = E( X 2 ) ( E( X ))2 (1.12) Var ( aX1 + bX2 + c) =

= a2 Var ( X1 ) + b2 Var ( X2 )
where the second moment E( X 2 ) is:
for any two independent
E( X 2 ) = Â x2 P( X = x ) (1.13) random variables X and
x Y and any real numbers
for discrete random variables, and: a, b and c.
Z
E( X 2 ) = x2 f X ( x )dx (1.14)
x
for continuous random variables.
4 probabilistic models for computer networks: tools and solved problems

Example 1. Find the expectation and variance of rolling a

die.

Solution The expectation and the variance for the experi-

ment of rolling a die are:
1 1 1 7
E( X ) = 1 + 2 + . . . + 6 = = 3.5
6 6 6 2
Now,
1 1 1 91
E( X 2 ) = 12 + 22 + . . . + 62 =
6 6 6 6
which yields:
✓ ◆2
91 7 35
Var ( X ) = E( X 2 ) ( E( X ))2 = = ⇡ 2.92
6 2 12

1.4 Some useful PDFs

The following list provides some useful PDFs along with

their expectation and variance values:

1. Continuous Uniform U ( a, b):

1
f X (x) = , axb (1.15)
b a
b a (b a )2
E( X ) = , Var ( X ) =
2 12
2. Discrete Uniform U ( a, b):
1
P( X = ki ) = , i = 1, . . . , n (1.16)
n
1 n 1 n
n i n iÂ
E( X ) = ki , Var ( X ) = (ki E( X ))2
=1 =1

3. Binomial Binom( N, p):

✓ ◆
N k
P( X = k) = p (1 p) N k
, k = 0, . . . , N (1.17)
k

E( X ) = N p, Var ( X ) = N p(1 p)
 probability review 5

4. Normal distribution N (µ, s2 ):

1 ( x µ )2
f X (x) = p e 2s2 , •x• (1.18)
2ps2

E( X ) = µ, Var ( X ) = s2

5. Exponential distribution Exp(l):

lt
f X ( x ) = le , t 0 (1.19)

1 1
E( X ) = , Var ( X ) =
l l2
6. Geometric distribution Geo ( p):
P ( X = k ) = (1 p) pk , k = 0, 1, 2, . . . (1.20)

p p
E( X ) = , Var ( X ) =
1 p (1 p )2
7. Poisson distribution Poiss(l):
lk l
P( X = k) = e , k = 0, 1, 2, . . . (1.21)
k!

E( X ) = l, Var ( X ) = l

1.5 Conditional and joint probability

Conditional probability refers to the probability of some

event X, given the occurrence of some other event Y, and
represented as:
P( X, Y )
P ( X |Y ) = (1.22)
P (Y )
where P( X, Y ) is referred to as the joint probability of ran-
dom variables X and Y. Such a joint probability accounts
for the case when both X and Y occur.
If random variables X and Y are independent, then:
P( X, Y ) = P( X ) P(Y ) (1.23)
The next example illustrates the concept of joint and
conditional probability.
6 probabilistic models for computer networks: tools and solved problems

Example 2. Consider the experiment of rolling two dif-

ferent dice. Let D1 and D2 refer to the random variables
that represent the result of each die separately, and let
X = D1 + D2 refer to the sum. Can you find the values
of P( X = 4) and P( X = 4| D1 = 1) and compare the
results?

Solution First, the sample space of X is SX = {2, 3, . . . , 12}.

There are 36 possible outcomes of the experiment ( D1 , D2 ),
listed in table 1.5.

Table 1.1: Possible out- (1,1) (1,2) (1,3) (1,4) (1,5) (1,6)
comes of ( D1 , D2 ) (2,1) (2,2) (2,3) (2,4) (2,5) (2,6)
(3,1) (3,2) (3,3) (3,4) (3,5) (3,6)
(4,1) (4,2) (4,3) (4,4) (4,5) (4,6)
(5,1) (5,2) (5,3) (5,4) (5,5) (5,6)
(6,1) (6,2) (6,3) (6,4) (6,5) (6,6)
For instance, the joint probability of having the outcome
(1, 3) is:
1 1 1
P( D1 = 1, D2 = 3) = P( D1 = 1) P( D2 = 3) = ⇥ =
6 6 36
since the results obtained from rolling the two dice are
independent.
However, the probability of scoring X = 4 is:
3
P ( X = 4) =
36
since this value takes into account the three possible out-
comes whose sum equals 4: (1, 3), (2, 2) and (3, 1) (see the
table above).
Next, consider the following probability value:
P( X = 4| D1 = 1)
This value should be different than P( X = 4) since some
information is already provided, i.e. the result of the first
die: D1 = 1. Indeed:
1 3
P( X = 4| D1 = 1) = 6= P( X = 4) =
6 36
 probability review 7

Essentially, given the fact that D1 scored a one, then D2

has to take the value: D2 = 3 in order to make X = 1 + 3 =
4, otherwise it would not be possible to score a 4. Then:

1
P( X = 4| D1 = 1) = P( D2 = 3) =
6
Essentially, the fact that D1 = 1 is known makes us
consider only the first column of the previous table, i.e.
results of type (1, ⇤). Then, D2 = 3 is just one case out of
six possible.
Alternatively, the previous result can also be obtained
from eq. 1.22:
P( X = 4, D1 = 1) 1/36 1
P( X = 4| D1 = 1) = = =
P( D1 = 1) 1/6 6
since both conditions, X = 4 and D1 = 1, are only met in
one case out of 36 possible (the case (1, 3)).

Example 3. Can you find the value of P( D1 = 1| X = 4)?

Solution This example requires to first consider all possi-

ble cases for X = 4 which are: (1, 3), (2, 2) and (3, 1). The
value P( D1 = 1| X = 4) requires to see how many of those
outcomes happen to have D1 = 1? The answer is clearly
only one: (1, 3) out of three possible. Thus:
1
P( D1 = 1| X = 4) =
3
Alternatively, the same result arises after applying
eq. 1.22:
P( D1 = 1, X = 4) 1/36 1
P( D1 = 1| X = 4) = = =
P ( X = 4) 3/36 3

1.6 Conditional expectation

The conditional expectation is the expected value of a ran-

dom variable with respect to a conditional probability dis-
8 probabilistic models for computer networks: tools and solved problems

tribution. Mathematically:

E ( X |Y = y ) = Â xP(X = x|Y = y) (1.24)

x

Example 4. Can you obtain E( X | D1 = 1) for the two-dice

example?

Solution We may compute E( X | D1 = 1) as:

E( X | D1 = 1) = Â xP(X = x| D1 = 1)
x
1 1 1
= 2 + 3 + . . . + 7 = 4.5
6 6 6
Alternatively, the previous value can also be obtained
from:
E( X | D1 = 1) = 1 + E( D2 ) = 1 + 3.5 = 4.5
since D1 = 1 is known.

Finally, it can be shown that:

E( X ) = E ( E( X |Y )) = Â E ( X |Y = y ) P (Y = y ) (1.25)
y

1.7 The Bayes’ theorem

The following two theorems are very useful in many prob-

abilistic problems:

Total probability theorem Given n mutually exclusive events

E1 , E2 , . . . , En whose probabilities sum to unity, then:
P( X ) = P( X | E1 ) P( E1 ) + . . . P( X | En ) P( En ) (1.26)

Bayes’ theorem The Bayes’ theorem shows the relationship

between two conditional probabilties:
P( X, Ek ) P( X | Ek ) P( Ek )
P( Ek | X ) = = (1.27)
P( X ) Âi P( X | Ei ) P( Ei )
 probability review 9

Example 5. Consider two urns, the first one contains two

white and eight black balls, and the second one contains
six white and four black balls. We toss a fair coin and then
we draw a ball from the first urn or the second urn de-
pending on whether the outcome of the coin was heads
or tails. The first question is: what is the probability that
a white ball is drawn? The second question is: what is the
conditional probability that the outcome of the toss was
heads given that a white ball was selected?

Solution The first question requires to make use of the

total probability theorem:

P(W ) = P(W |U1 ) P(U1 ) + P(W |U2 ) P(U2 )

2 1 6 1 2
= ⇥ + ⇥ =
2+8 2 6+4 2 5
The solution to the second question arises after applying
the Bayes’ theorem:
2 1
P(W |U1 ) P(U1 ) 2+8 2 1
P(U1 |W ) = = 2
=
P (W ) 5
4

1.8 The Central Limit Theorem

Consider the following set of random variables: { X1 , X2 , . . . , Xn },

each of which happens to be independent and identically
distributed (i.i.d.) with finite mean µ and variance s2 . We
know that the sum:
S n = X1 + . . . + X n
has mean nµ and variance ns2 .
The Central Limit Theorem states that the random vari-
able:
Sn nµ
Zn = p
s n
10 probabilistic models for computer networks: tools and solved problems

approximates the standard normal distribution N (0, 1) for

large n. In particular:
Z z
Sn nµ 1 u2
lim P( p  z) = p e 2 du (1.28)
n!• s n 2p •

Figure 1.4: CLT in action: 0.18

Hist.
0.08
Hist.
Norm. fit
the sum of two and ten
0.16 Norm. fit
0.07

0.14
dice 0.12
0.06

0.05
0.1
0.04
0.08

0.03
0.06

0.02
0.04

0.02 0.01

0 0
2 3 4 5 6 7 8 9 10 11 12 0 10 20 30 40 50 60 70
S2 S10

(a) Normal fit for S2 (b) Normal fit for S10

Example 6. Let Sn denote the result of rolling n and adding

up their results. Thanks to the Central Limit Theorem, we
know that the PDF of Sn approaches a Normal distribution
N ( 42 35
12 n, 12 n ).
Figs. 1.4(a) and 1.4(b) show examples of the the appli-
cability of the CLT for the sum of two and ten dice respec-
tively.

1.9 Further problems

Problem 1. Consider a computer that may send Data

frames or Acknowledgements with equal probability.
Compute the probability that the first time that two data
frames are transmitted consecutively occurs on the third
and fourth frames.

Solution This experiment requires that the third and

fourth frame are Data frames. Also, the second frame
must be an Ack, otherwise we would have the first data
frames on the second and third frames, rather than third
 probability review 11

and fourth. Finally, the first frame can either be a Data

frame or Ack. Hence:
P(⇤, Ack, Data, Data) = 1 ⇥ 0.5 ⇥ 0.5 ⇥ 0.5 = 0.125

Problem 2. Consider a wireless network with 13 radio

channels. In this network, two computers decide to use
any of these channels for communicating with an access
point. There is a collision if the same channel is used by
any two computers. a) Compute the probability of having a
collision when the network comprises only two computers.
b) Find the collision probability for three computers.

Solution a) Let CO2 refer to the random variable that

takes value 1 if there is a collision or 0 otherwise, i.e.
SCO2 = {0, 1}; and let Ci refer to the channel used by the
i-th station, where SCi = {1, . . . , 13}. There is a collision if
both stations happen to select the same channel, that is:
P(CO2 = 1) = P(C1 = 1, C2 = 1) + P(C1 = 2, C2 = 2) + . . .
+ P(C1 = 13, C2 = 13)
1 1 1 1 1 1
= +...+ = 13 ⇥ 2 =
13 13 13 13 13 13
Also, this probability may be obtained just by counting
the number of cases that produce a collision, and divide
this number over the total number of cases. There are thir-
teen collision cases:
(1, 1), (2, 2), . . . , (13, 13)
over a total number of cases of: 13 ⇥ 13 = 169. Thus:

no. possible cases 13 1

P(CO2 = 1) = = 2 =
no. total cases 13 13
Similarly, there are 133 = 2197 possible cases when we
have three stations instead of two.
Now, we need to compute the number of collision cases.
For channel no. 1, we have three sets of collision cases:
12 probabilistic models for computer networks: tools and solved problems

(1, 1, ⇤), (1, ⇤, 1) and (⇤, 1, 1) where ⇤ may take any value
from 2 to 13. That is a total of 12 ⇥ 3 cases. The same rea-
soning applies for channel 2: (2, 2, ⇤), (2, ⇤, 2) and (⇤, 2, 2)
yielding another set of 12 ⇥ 3 set of collision cases. Same
applies for channels 3, . . . , 13. Thus we have: 13 ⇥ 12 ⇥ 3
cases where any two stations select the same two channels.
Finally, we also need to include the cases where all three
stations happen to select the same channel:
(1, 1, 1), (2, 2, 2), . . . , (13, 13, 13)
which is another set of 13 cases. Thus:
13 ⇥ 12 ⇥ 3 + 13 37
P(CO3 = 1) = =
133 169
Alternatively, the same result arises with the following
reasoning: There is no collision if the second station does
not select the same channel as the first station, and the
third station does not select the same channel as station 1
and 2. Hence, the first station has 13 channels to choose
from, the second station only has 12 channels to choose
from (one is already occupied by station 1) while the third
station only has 11 channels to choose from (two channels
occupied by stations 1 and 2). Thus, the total number of
non-collision cases is: 13 ⇥ 12 ⇥ 11 = 1716. Hence:
13 ⇥ 12 ⇥ 11 37
P(CO3 = 1) = 1 =
133 169

Problem 3. An Ethernet switch has got 10 computers and

5 servers attached to it, each of them transmits packets to
the Internet. The computers transmit 64-byte frames (TCP
ACKs) with probability 0.6 and 1518-byte frames (Data
frames) with probability 0.4, while the servers transmit 64-
byte frames with probability 0.1 and 1518-byte frames with
probability 0.9. Find the probability that a given selected
frame is of size 64 bytes. What is the probability that a
server has generated this frame?
 probability review 13

Solution Let random variable FS denote the size of a given

frame, with the following possible values SFS = {64, 1518}
bytes. Now, let E refer to the the random variable that
denotes the source of a given frame, either Computer or
Server. Then the probability of observing a 64-byte frame
equals (total probability theorem):

P( FS = 64) = P( FS = 64| E = C ) P( E = C ) + P( FS = 64| E = S) P( E = S)

10 5 13
= 0.6 ⇥ + 0.1 ⇥ = = 0.43
10 + 5 10 + 5 30

Then, 43.3% of the frames in the network are short while

the other 56.7% are long 1518-byte frames.
Now, a given short packet actually comes from a server
with probability (Bayes’ theorem):

P( FS = 64, E = S) P( FS = 64| E = S) P( E = S)
P( E = S| FS = 64) = =
P( FS = 64) P( FS = 64)
0.1 ⇥ (5/15) 1
= =
13/30 13

App. Packet size dist. %

Skype U(50,150) 5%
Problem 4. The amount of traffic received by a computer is P2P U(1000,1500) 60%
measured to follow Table 1.2. Web Exp(1/1000) 25%
In this table, the first column gives the application type email N(800,100) 10%
Table 1.2: Packet size dis-
of the packets, the second column refers to the packet size tribution per application
distribution for that application and the third column in-
dicates the percentage of traffic over the total for such an
application. Obtain the average packet size.

Solution This problem requires to make use of the condi-

tional expectation equations. Let PS refer to the random
variable denoting packet size, and let A refer to the appli-
cation type. Then:
14 probabilistic models for computer networks: tools and solved problems

E( PS) = E( E( PS| A)) = Â E( PS| A) P( A)

A
150 + 50 1500 + 1000
= 0.05 · + 0.6 ·
2 2
+ 0.25 · 1000 + 0.1 · 800
= 1085 bytes
which gives the average packet size weighted per applica-
tion.

Problem 5. A wireless network uses Stop-and-Wait ARQ

to recover lost frames in a lossy media. Such a lossy media
has a frame loss probability of 0.1, and Maximum Transfer
Unit (MTU) of 1000-byte packets. Assume, that a given
user wants to download a file of 10000 bytes. Compute
(a) the probability that no frame needs retransmission, (b)
exactly two frames need retransmission. Compute also (c)
the average number of frames that need retransmission.

Solution (a) First a retransmission is needed in Stop-and-

Wait ARQ if either a data frame or its Ack (or both) are
lost. This occurs with probability:
pretx = P( Data = Lost, Ack = Lost)
+ P( Data = Ok, Ack = Lost)
+ P( Data = Lost, Ack = Ok)
= 0.1 · 0.1 + 0.9 · 0.1 + 0.1 · 0.9 = 0.19
Also, the number of data frames to be transmitted is:
file size
N= = 10
Function binopdf from MTU
Matlab is very useful The number of retransmissions required can be mod-
to deal with binomial elled with the binomial distribution. That is, k frames re-
distributions: quire one or more retransmissions with probability:
✓ ◆
binopdf(k,N,p) N k
P(k) = pretx (1 pretx )10 k
k
 probability review 15

Hence, no frame needs retransmission with probability:

✓ ◆
10 0
P ( k = 0) = pretx (1 pretx )10 0 = (1 0.19)10 = 0.1216
0
Similarly, exactly two frames need retransmission with
probability:
✓ ◆
10 2
P ( k = 2) = pretx (1 pretx )8 = 0.301
2
In this case:
The average number of retransmissions needed is:
binopdf(2,10,0.19)

E(k ) = N p = 1.9 frames

Problem 6. We are interested in using slotted Aloha with

service differentiation to control medium access with dif-
ferent priority levels. The difference with conventional
slotted Aloha is that high-priority users may transmit on a
slot with higher probability than low-priority users.
Consider a network with two stations only: One is of
high-priority and the other one is of low priority, whereby
the transmission probability for high-priority stations is
double than that for low-priority stations. a) Obtain the
transmission probability for each type of station that max-
imises the probability of successfully transmitting any data
on a time slot. b) Formulate the problem for the case of
three stations: one of high and two of low priority.

Solution Let Si , i = 1, 2 denote the state of the i-th station:

Tx, Idle.
There is a successful transmission on a timeslot if only
one of the two stations transmit during a time-slot. This
occurs with probability:

Psuccess = P(S1 = Idle, S2 = Tx ) + P(S1 = Tx, S2 = Idle)

= (1 2p) p + 2p(1 p) = 3p 4p2
16 probabilistic models for computer networks: tools and solved problems

To maximise this quantity, we need to differenciate with

respect to p and set that value to zero:
dPsuccess 3
= 0 = 3 8p ) p =
dp 8
We observe that this is a maximum since differentiating
again gives a negative value for p = 3/8.
When p = 3/8, a successful time slot occurs with proba-
bility:
3 32
Psuccess = 3 4 2 = 0.5625
Function roots from 8 8
Matlab allows to find the In the case of two low-priority and one high-priority
polynomial roots. In this stations, we have:
✓ ◆ ✓ ◆
case, we should run: 2 1 2 0
2 1
Psuccess = (1 2p) p (1 p ) + 2p p (1 p )2 0
roots([18,-20,4]) 1 0
= 4p 10p2 + 6p3
to obtain the result.
We can further use again, maximising this quantity requires:
Matlab to plot: dPsuccess
= 0 = 4 20p + 18p2 ) p = {0.2616, 0.8495}
Psuccess = 4p 10p2 + 6p3 dp
Only the first value maximises Psuccess = 0.4695, since the
using (see Fig. 1.5):
second value gives a minimum, rather than a maximum.
p=linspace(0,1,100);

plot(p,4*p-10*p.^2+6*p.^3)
Problem 7. A group of 2N stations are connected through
0.4 a slotted Aloha system, where the capture effect is used to
0.3
solve collisions. This means that if several stations transmit
Psuccess

0.2
0.1 at the same time, only the signal from the station with the
0
−0.1
highest power level is valid at the receiver, the other ones
0 0.2 0.4 0.6 0.8 1 are not.
p
Now, consider a network with 2N stations, divided into
Figure 1.5: Psuccess plotted
with Matlab for problem N power classes with two stations per class. Assume that
6. each station, transmits on a time-slot with probability p.
Obtain: a) The probability of receiving data from any of the
two stations of the highest power level. b) The probability
of receiving data from any of the two stations of the lowest
power level. c) The probability of a successful transmission
over a timeslot, when N ! •.
 probability review 17

Solution In the first case, we receive data from any of the

two highest-power stations if only one of them transmits a
packet:
✓ ◆
2 1
p (1 p)2 1 = 2p(1 p)
1
In order to receive data from any of the lowest-power
stations, it is necessary that none of the other 2N 2 sta-
tions transmit any packet, and only one of the two lowest-
power stations transmit a packet. This occurs with proba-
bility:
✓ ◆ ✓ ◆
2N 2 0 2 1
p (1 p)2N 2 ⇥ p (1 p)2 1 = 2p(1 p)2N 1
0 1
Basically, there is a collision if two stations of the same
power group decide to transmit on the same time-slot,
while the others from the upper power group do not trans-
mit any packet. So, there is a successful transmission if
one station of the first group decides to transmit a packet,
or if one station of the second group decides to transmit
(while the two stations of the first group do not), and so
on. Hence:

Psuccess = 2p(1 p ) + (1 p)2 2p(1 p) + (1 p)4 2p(1 p) + . . .

•
n 2(1 p )
= 2p(1 p) Â (1 p )2 =
2 p
n =0

Problem 8. A WiFi station has two modulation formats

to transmit its packets: Mod1 at 1 Mbit/s and Mod2 at
11 Mbit/s. When using Mod2, the frame error probabil-
ity experienced in the channel is 20%. However, Mod1 is
slower but more reliable: the frame error probability is only
3%. An external observer measures that about 12.35% of
the frames have been lost. Compute: (a) The percentage
of frames transmitted using Mod1 and Mod2; and (2) the
18 probabilistic models for computer networks: tools and solved problems

percentage of successfully received frames that were trans-

mitted at Mod2.

Solution First, we know that the two modulation formats

have different error probabilities, i.e.: P( Err | Mod1) = 0.03
and P( Err | Mod2) = 0.2.
Thus, from the total probability theorem, we know that:
P( Error ) = P( Err | Mod1) P( Mod1) + P( Err | Mod2) P( Mod2)
= 0.03 · P( Mod1) + 0.2 · P( Mod2) = 0.1235
However, P( Mod1) = 1 P( Mod2) since the station uses
one modulation or the other. Then:
P( Err ) = 0.03 · P( Mod1) + 0.2 · (1 P( Mod1)) = 0.1235
Therefore:
0.2 0.1235
P( Mod1) = = 0.45
0.2 0.03
and P( Mod2) = 0.55.
Thanks to the Bayes’ theorem, we obtain the percentage
of successful packets that were transmitted using Mod2:

P( NoErr | Mod2) P( Mod2)

P( Mod2| NoErr ) =
P( NoErr )
0.8 · 0.55
= = 0.502
0.8 · 0.55 + 0.97 · 0.45

Problem 9. A server in a company has got four processors

(four cores) to serve users’ jobs. Consider that only three
users can log into the server. Each user may offer zero, one
or two jobs, with the following probabilities: he offers one
job with probability 2p, and two jobs with probability p. (a)
Find the average number of used processors when p = 0.1.
Find also the variance. (b) Find the value of p that achieves
100% utilisation of the four processors on average. (c) For
the value of p computed in the previous case, obtain the
probability of having more jobs than processors.
 probability review 19

Solution First, let Ui refer to the number of jobs offered by

the i-th user, with SUi = {0, 1, 2}. Clearly, the total number
of jobs offered is:
U = U1 + U2 + U3
The first and second moments of U1 are:
E(U1 ) = 0 · (1 p 2p) + 1 · 2p + 2 · p = 4p = 0.4
and:
E(U12 ) = 02 · (1 p 2p) + 12 · 2p + 22 · p = 6p = 0.6
Hence, the variance is:
Var (U1 ) = E(U12 ) ( E(U1 ))2 = 6p 16p2 = 0.6 0.16 = 0.44
Finally, the mean and variance of U are:
E(U ) = E(U1 ) + E(U2 ) + E(U3 ) = 12p = 1.2
Var (U ) = Var (U1 ) + Var (U2 ) + Var (U3 ) = 1.32
since the number of jobs offered by each user are indepen-
dent from one another.
The value of p that achieves 100% utilisation of the four
processors requires to solve:
1
E(U ) = 12p = 4 ) p =
3
In general, we have more jobs than servers in the follow-
ing set of cases for (U1 , U2 , U3 ): (2, 2, 1), (2, 1, 2), (1, 2, 2)
and (2, 2, 2), that is, the cases when one user offers only
one job but the other two offer two jobs, or when all three
users offer two jobs each. The first case occurs with proba-
bility:
6
pp(2p) + p(2p) p + (2p) pp = 6p3 =
27
and case (2, 2, 2) happens with probability:
1
ppp = p3 =
27
In total, the probability that more jobs are offered than
processors:
7 1
when p =
27 3
2
The exponential random variable

2.1 Definition

Let X be an exponentially-distributed random variable,

characterised by some parameter l > 0. The Probability
Density Function (PDF) of X follows: 2
1.8
Exponential PDF

λ=1
λ=2
lt
f X (t) = le , t 0 (2.1) 1.6
1.4
1.2

and has the appearance of Fig. 2.1. As shown, the larger fX(t)
1
0.8

the value of l, the faster the tail decays. 0.6

0.4

The Cumulative Distribution Function (CDF) of X is 0.2

0
0 0.5 1 1.5 2

given by:
t

Z t Figure 2.1: Exponential

FX (t) = P( X  t) = le lt
dt = 1 e lt
, t 0 (2.2) PDF.
t =0
and the Complementary Cumulative Distribution Function
(CCDF) or Survival function is then:
lt
SX (t) = P( X > t) = 1 FX (t) = e , t 0 (2.3)
The mean and variance
Z •
can be observed to follow:
1
E( X ) = t f X (t)dt = (2.4)
0 l
Z •
1
Var ( X ) = (t E( X ))2 f X (t)dt = 2 (2.5)
0 l

Example 1. Consider a light bulb whose duration is expo-

nentially distributed with parameter l = 2 failures/year.
Obtain the average duration of this light bulb and its stan-
dard deviation.
22 probabilistic models for computer networks: tools and solved problems

Solution The average duration of this light bulb is:

1
E( X ) = = 0.5 year
l
and its standard deviation is:
q
1
Std( X ) = Var ( X ) = = 0.5 year
l

2.2 Simulation and estimation

Consider we have an exponentially-distributed random

variable X for some parameter l. Then, we may obtain a
number of observations drawn from X as:
ln ui
oi = , i = 1, 2, . . . (2.6)
l
where the ui values are uniformly-distributed random
numbers within the interval (0, 1), i.e. ui ⇠ U (0, 1). The
above equation transforms uniformly-distributed random
numbers into exponentially-distributed random values
with parameter l.
The next figure shows 40 observations drawn from an
exponential random variable with parameter l = 2. As
shown, most of the values lie within the range t 2 [0, 1].
This interval is exactly the mean plus/minus the standard
deviation, i.e. E( X ) ± Std( X ).

Figure 2.2: Random 1

Random samples from X∼ exp(λ)

samples drawn from

exp(l = 2) 0

−1
0 0.5 1 1.5 2 2.5 3 3.5
t

Now, consider a set of observations O = {o1 , . . . , o N }

drawn from an exponential random variable X whose pa-
rameter is unknown. The easiest way to estimate its pa-
rameter is via the first moment estimation. Essentially, we
 the exponential random variable 23

know that the expection of X equals E( X ) = l 1 . Thus, we

may obtain the average value of the observations as:
N
1
X̄ =
N Â oi (2.7)
i =1

and approximate parameter l as:

! 1
N
1 1
l̂ =
X̄
=
N Â oi (2.8)
i =1

Now, consider that we are asked to determine whether

or not some set of observations O = {o1 , . . . , o N } where
drawn from some exponentially-distributed random vari-
able X. The best way to proceed here is to estimate the
CCDF (or survival function) of X based on such obser-
vations and see whether or not its logarithm follows a
straight line. This is based on the fact that the survival
function SX (t) of an exponential random variable follows
eq. 2.3, hence its logarithm:

ln SX (t) = lt (2.9)

follows a straight line with slope l.

Fig. 2.2 shows an example of ln SX (t) for a set of ob-
servations that follow the exponential distribution with
parameter l = 2.

0
10 Figure 2.3: Log-survival
log(SX (t)) for an expo-
nential PDF
−1
10
log(SX(t))

−2
10

−3
10
0 0.5 1 1.5 2 2.5 3 3.5 4
t
24 probabilistic models for computer networks: tools and solved problems

2.3 The memoryless property

The exponential is the only distribution that satisfies the

memoryless property which, in mathematical words, states
that:
P ( X > t + s | X > s ) = P ( X > t ), s, t > 0 (2.10)
To prove this property, it is important to observe that:
P( X > t + s, X > s)
P( X > t + s| X > s) =
P( X > s)
P( X > t + s) e l(t+s)
= =
P( X > s) e ls
which turns out to be the same as:
lt
P( X > t) = e
The memoryless property
is only satisfied by the The implications of this property are key in under-
exponential distribution; standing Poisson processes and Continuous-Time Markov
all other distributions
Chains, which will be explored in subsequent chapters.
have memory, i.e.:
Looking deeper in eq. 2.10, we observe that the memo-
P( X > t + s| X > s) ryless property states that the past s units of time are not
6= P( X > t) important in predicting the future behaviour of our expo-
nential random variable X.
Essentially, consider the example of the duration of a
light bulb characterised by an exponential distribution with
parameter l = 5 failures/year. Then, we know that the av-
erage duration of this light bulb is E( X ) = l 1 = 0.2 year
(or 73 days). Now, consider that we switch on the light on
day 0, and we observe that the light bulb works perfectly
for s = 20 days. If we are asked about the expected du-
ration of this light bulb from day 20 onwards, the answer
should be:
My light bulb should last for another 73 days on average!!!
This conclusion seems counter-intuitive. One would
expect that, since the light bulb has already “consumed” 20
days, the remaining life expectancy should be 73 20 = 53
 the exponential random variable 25

days. This might be true for probability distributions with

memory, but is wrong for the exponential distribution due
to the memoryless property. The next example proves this
result. Thanks to the memory-
less property, information
about the past of an
Example 2. Consider a light bulb whose duration is char- exponential random vari-
acterised by an exponential distribution with parameter able X is not relevant in
its future behaviour.
l = 5 failures/year. Consider that the light bulb has al-
ready survived T = 20 days. Obtain its average duration
assuming X > T, i.e. E( X | X > T ).

Solution E( X | X > T ) can be obtained as follows:

Z •
E( X | X > T ) = t f X (t | X > T )dt
T

This requires to obtain the conditional PDF first:

P( X = t, X > T ) le lt l(t T )
f X (X = t|X > T) = = = le , t T
P( X > T ) e lT
Hence, the integral above becomes:
Z •
l(t T )
E( X | X > T ) = tle dt
T

which, after some calculus, yields:

1
E( X | X > T ) = T + days
l
Hence, the expected total duration of the light bulb is
T + E( X ) = 20 + 73 = 93 days, that is, another E( X ) = 73
days more than the already consumed T = 20 days.

In conclusion, the memoryless property states that the

previously observed duration of the light bulb is not im-
portant in determining its future behaviour. The “remain-
ing” lifetime of the light bulb is again exponentially dis-
tributed with parameter l.
26 probabilistic models for computer networks: tools and solved problems

2.4 Minimum of several exponential random variables

Another important property of the exponential distribu-

tion concerns the minimum of several exponential random
variables:

Minumum Let X1 , . . . , X N be independent exponential

random variables charecterised by parameters l1 , . . . , l N
respectively. Then, the minimum of them, denoted as:
Xm = min{ X1 , . . . , X N }
is also exponentially distributed with parameter:
l m = l1 + . . . + l N

Essentially:
P( Xm > t) = P( X1 > t, X2 > t, . . . , X N > t) (2.11)
which implies that all random variables Xi have to be
greater than t to satisfy that the minimum of them is also
greater than t. Both requirements are equivalent. Then:

P ( X m > t ) = P ( X1 > t ) P ( X2 > t ) · · · P ( X N > t ) = e ÂiN=1 li t

(2.12)
which proves that the random variable Xm is also exponen-
tially distributed with parameter lm = ÂiN=1 li .

Example 3. Consider a car with three main components

subject of failure: its engine, its electricity system, and its
air conditioning, each characterised by an exponential ran-
dom variable with parameters: le = 1 failure per decade,
les = 2 failures per decade and l ac = 5 failures per decade
respectively. Find the probability of having one failure of
any type in its first year.

Solution Let Xm denote the first failure of the car. This

first failure is the minimum of the three exponential ran-
dom variables that characterise the car’s engine, electricity
 the exponential random variable 27

system and air conditioning. Thus, Xm is exponentially

distributed with parameter lm :
lm = le + les + l ac = 1 + 2 + 5 = 8 failure/decade,
or 0.8 failure/year. Hence, we have the first failure in less
than a year with probability:
0.8⇥1
P( Xm  1 year) = 1 e = 0.55
It is finally worth remarking that the time units must
match with parameter l. If instead of using years as time
units we use decades, then the correct calculus would be:
8⇥0.1
P( Xm  0.1 decade) = 1 e = 0.55
where we have used parameter l = 8 failure/decade in
order to match the 0.1 decade time value.

2.5 Comparison of several exponential random variables

Comparison Consider two independent exponential random

variables X1 and X2 characterised by parameters l1 and
l2 . Then:
l1
P ( X1 < X2 ) = (2.13)
l1 + l2
This result arises after applyng the total probability theo-
rem for all possible values of X2 :

Z •
P ( X1 < X2 ) = P( X1 < X2 = t| X2 = t) P( X2 = t)dt
0
Z •⇣ ⌘ l1
= 1 e l1 t l2 e l2 t dt =
0 l1 + l2
This property allows to compare exponential random
variables very easily, just as the ratio of their rate with
respect to the total.

Example 4. Consider the same car as in the previous ex-

ample. Find the probability that the engine fails before the
28 probabilistic models for computer networks: tools and solved problems

other two components in the car. Find also the probabil-

ity that the engine fails twice in a row, that means, first
and second failure also occurs in the engine (the engine is
assumed to be repaired after the first failure).

Solution The engine fails before the other two components

with probability:
le 1 1
P( Xe  min{ Xes , Xac }) = = =
le + (les + l ac ) 1+2+5 8
Here, we have also applied the property that the minimum
of two exponential random variables is also exponentially
distributed.
Next, the engine fails twice in a row with probability:
1 1
8 ⇥ 8 . Essentially, once the first failure occurs, the system
renews itself and starts all over again, thanks to the mem-
oryless property of the exponential distribution. In other
words, after the first engine failure and its repair, the next
failure does not take into account the history or past of the
air conditioning or electricity system (memorylessness).

The properties of minimum and comparison of expo-

nential random variables have further applications in the
Poisson process and Continuous-Time Markov Chains.

2.6 Further problems

Problem 1. Prove that the uniform distribution does not

have the memoryless property.

Solution This requires to check whether or not:

P( X > t + s|t > s) = P( X > t)

holds for the uniform distribution U (0, a), i.e.:

1
f X (t) = , 0ta
a
Exploring the Variety of Random
Documents with Different Content
year to be good in, as little Christian says, a new year to live and learn in. It
is true that, perhaps, you may not see its end; but, nevertheless, it is the
beginning of a new year with many opportunities, both of doing and
receiving good, and therefore we should be glad, and we should ask God to
make us His faithful servants, loving Him and keeping His commandments
all through this year, and if God does that you may be sure this will be a
very happy new year to us all. Well, Halbert,” he continued, turning to his
son, who was back again by Aunt Christian’s side, “has little Christian
satisfied you?”
Halbert’s face and conscience were both quite cleared; it was right to be
glad on a new year’s day, and he got a promise that that night he should
hear some of the many things which had happened on former new years’
days, and had made that day a special anniversary in the family; and
besides, the relation of these things was to be committed to Aunt Christian,
therefore Halbert was quite satisfied. And then the seniors closed round the
fireside, and all the children—with the exception of Halbert Melville and
Mary Hamilton, the eldest of the two families, who hang by Aunt Christian
still—sought more active amusement in the farther corners of the room, and
recollections of those bygone years became the long lingered on subject
with Halbert, Charles, Christian, and the two Marys; and they looked back
with half-wondering gaze upon the past, as men look through the wondrous
glass of science on the clear outline of some far distant shore, of which the
human dwellers, the fears and hopes, the loves and sorrows, which people
the farther sides of the blue slopes that yet linger in their view, have all
faded from their retiring vision.
But then comes a distant shout from the lobby into which some of the
children have strayed in their play, of “Uncle James! Uncle James!” and
here he is. Older, of course, yet looking much as he looked in the old times;
though we must whisper that the bridegroom whom we saw some fourteen
or fifteen years ago at the commencement of this story, has now, at its
conclusion, become a portly gentleman; in good sooth, most
unsentimentally stout, and with a look of comfort and competence about
him, which speaks in tones most audibly, of worldly success and prosperity.
A good man, too, and a pleasant, he is, with the milk of human kindness
abounding in his heart; as such Mr. James Melville is universally
considered and honoured, though with scarcely so large a heart as his
brother the minister, nor so well mated. It is true, Mrs. James, since she
found out who her friend of ten years ago was; and Mary’s reasons for
rejecting what seemed so good a match, and the failure, the utter failure of
her party on that new year’s night in consequence; has grown wonderfully
careful, and begins to discover that there are pleasanter things in life, than
the collecting together a dozen or two of people to be entertained or wearied
according to their respective inclinations, and her fireside has grown a
much more cheerful one always, though for a few nights in the year less
brilliant than heretofore; and her husband’s quotations of “Christian” have
grown less disagreeable to her ears, though still she sometimes resents the
superiority which everybody accords to her. James is always welcomed in
his brother Halbert’s house, and never more warmly than on New Year’s
night; for Elizabeth does not accompany him on these annual occasions;
and even that loving circle feel relieved by her absence at such a time, for
the conversation generally runs upon certain remembrances which she
would not like to hear; and which none of them would like to mention in
her presence. So James sits down and joins them for awhile in their
recalling of the past; and little Halbert Melville gazes at his father in open-
mouthed astonishment, as he hears him speak of being the cause of
unhappiness and sorrow to Aunt Christian and Aunt Mary, and to Uncles
James and Robert, and his grave old grandfather who died two years ago.
His father—and Halbert would have defied anybody but that father’s self.
Yes! even Aunt Christian, if she had said such words as these—his father
cause unhappiness and sorrow to anybody!—his father, whom old Ailie,
still a hale and vigorous old woman, and chief of Christian’s household, and
prima donna in Mary Melville’s nursery, had told him was always as kind
and good to everybody all through his life as he was now! Halbert could not
believe it possible. And little Mary Hamilton’s eyes waxed larger and
larger, in amazement, as Aunt Christian spoke of her mother—her mother
whom she had never seen without a smile on her face, being at that
infinitely remote period before any of them were born, most unhappy
herself; yes, very unhappy! Mary would have denied it aloud, but that she
had too much faith in Aunt Christian’s infallibility, to doubt for an instant
even her word. This night was a night of wonders to these two listening
children.
But the time passed on, and Uncle James—while yet the other little ones
were engaged in a merry game, chasing each other throughout all the house,
from the glowing kitchen, clean and bright, up to the nursery where old
Ailie presided in full state and glory—must go. Elizabeth was unwell; and
he felt it was not seemly to be from home, loth and reluctant as he was to
leave that fireside and its loving circle. So Uncle James prepared to go
home; and down rushed again the whole merry band, deserted Ailie, even in
the midst of one of her old-world stories, to bid him good-night; and thus
environed by the little host with shouts as loud as had welcomed his arrival,
Uncle James went away home.

CHAPTER II.

Men rail upon the Change!

* * * * *
But think they as they speak?
Thou softener of earth’s pain,
Oh Change! sweet gift of the Infinite to the weak,
We hail alike thy sunshine and thy rain;
Awe dwells supreme in yon eternal light,
Horror in misery’s doom;
But frail humanity dares breathe, when bright
Thy tremulous radiance mingles with the gloom.—Y.S.P.

NCLE JAMES has just gone, and the group of elders in the
parlour are just drawing their chairs closer together to fill up the
gap which his departure has made, when they hear a hasty knock
at the door; a hasty, imperative summons, as if from urgent need
that would not be denied access, and a dripping messenger stands on the
threshold—for the cold rain of winter falls heavily without—begging that
Mr. Melville would go with him to see a dying man, a stranger who has
taken up his residence for the last few weeks at a small inn in the
neighbourhood, and was now, apparently, on the very brink of death, and in
a dreadful state of mind. The calls of the sick and dying were as God’s
special commands to Halbert; and he rose at once to accompany the
messenger, though the faces of his wife and sisters twain, darkened with
care as he did so. It was very hard that he should be called away from them
on this especial night; and when he firmly declared he would go, Mary
whispered to Charles to go with him, and to bring him soon back. The two
brothers went away through the storm, and the sisters drew closer to each
other round the fire, as the gentlemen left them; then Mrs. Melville told the
others how anxious she always was when her husband was called out in this
way; how he might be exposed to infection in his visiting of the sick so
assiduously as he did; and how, for his health’s sake, she could almost wish
he were less faithful and steady in the discharge of these his duties: and
Mary looked at her in alarm as she spoke, and turned pale, and half
upbraided herself for having unnecessarily exposed Charles, though a more
generous feeling speedily suppressed her momentary selfishness. But
Christian was by, and when was selfishness of thought, or an unbelieving
fear harboured in Christian’s gentle presence?
“Mary! Mary!” she exclaimed, as she turned from one to the other, “are
you afraid to trust them in the hands of your Father? They are but doing
what is their duty, and He will shield His own from all evil. Would you
have your husband, Mary Melville, like these ministers whose whole work
is their sermons—alas! there are many such—and who never try, whether
visiting the sick and dying, or the vicious and criminal, would not advance
their Master’s cause as well—would you that, rather than Halbert’s going
forth as he has done to-night?”
“No, no; but it is terrible for me to think that he is exposed to all kinds of
contagion; that he must go to fevers, and plagues, and diseases that I cannot
name nor number, and run continually such fearful risks,” said Mary,
energetically.
“Our Father who is in Heaven, will protect him,” said Christian,
solemnly. “I have heard of a minister in London, who never for years ever
thinks of seeing after his own people in their own homes; it is too much
labour, forsooth, he is only their preacher, not their pastor; and though he
sends—Reverend Doctor that he is—his deacons and such like to visit; it’s
seldom that himself ever goes to a poor sick bed, and as to his trying to
reclaim the vicious, there is not on his individual part the least attempt or
effort. Now, Mary, would you have Halbert such a man as that?”
“I would rather see him lying under the direfullest contagion. I would
rather that he was stricken by the Lord’s own hand, than that it should be
said of Halbert Melville that he flinched in the least degree from the work
which the Lord has laid upon him,” returned Mary, proudly elevating her
matronly form to its full height, with a dignity that gladdened Christian’s
heart.
“Yet that man in London will be well spoken of,” said Mary Hamilton,
“and our Halbert unknown. No matter: the time will come when Halbert
will be acknowledged openly; and now, Christian, I feel assured and
pleased that Charles went out with Halbert.”
“And you may, when they went on such an errand,” said Christian;
“but”—and she continued briskly, as if to dispel the little gloom which had
fallen upon them, and resuming the conversation, which had been broken
off on the departure of the gentlemen—“but Robert writes me, that he is
very comfortably settled, and likes his new residence well.”
“I am sorry,” said Mrs. Melville, after a pause, during which her
agitation had gradually subsided, “I am sorry that I saw so little of Robert.
He and I are almost strangers to each other.”
“Not strangers, Mary, while so nearly connected,” said Christian, kindly.
“Moreover, Robert gives me several very intelligible hints about a young
lady in your uncle’s family to whom you introduced him.”
“Indeed!” exclaimed Mrs. Melville, “no doubt he means my cousin
Helen. Oh, I am very glad of that. Your brothers are too good, Christian, to
be thrown away on cold-hearted, calculating people, who only look at
money and money’s worth——” and as the words fell from her lips, she
stopped and blushed, and hesitated, for Mrs. James flashed upon her mind,
and the comparison seemed invidious.
“You are quite right, Mary,” said the other Mary, smiling; “and if Robert
be as fortunate as Halbert has been, we shall be a happy family indeed.”
Did Christian’s brow grow dark with selfish sorrow, as she listened to
these mutual congratulations? Nay, that had been a strange mood of
Christian’s mind in which self was uppermost, or indeed near the surface at
all; and her whole soul rejoiced within her in sympathetic gladness. Nor,
though they were happy in the full realisation of their early expectations,
did she hold herself less blessed; for Christian bore about with her, in her
heart of hearts, the holy memory of the dead, and in her hours of stillest
solitude felt not herself alone. An angel voice breathed about her in
whispering tenderness when she turned over the hallowed leaves of yon old
Bible; and when the glorious light of sunset fell on her treasured picture, it
seemed, in her glistening eyes, to light it up with smiles and gladness; and
the time is gliding on gently and silently, day upon day falling like leaves in
autumn, till the gates of yon far celestial city, gleaming through the mists of
imperfect mortal vision, shall open to her humble footsteps, and the beloved
of old welcome her to that everlasting reunion; and therefore can Christian
rejoice, as well on her own account, as in ready sympathy with the joyful
spirits round about her.
But the present evening wore gradually away, and the children became
heavy, weary, and sleepy, and the youngest of all fairly fell asleep; and Mrs.
Melville looked at her watch anxiously, and Mary said she could not wait
for Charles, but must go home; but here again Christian interposed. The
little Melvilles and Hamiltons had slept under the same roof before now,
and being too far gone in weariness to have joined in their domestic
worship, even had the elders been ready to engage in it, were taken off by
twos and threes indiscriminately to their respective chambers; and the three
sisters are left alone once more, maintaining, by fits and starts, a
conversation that showed how their thoughts wandered; and, in this dreary
interval of waiting for the home-coming of Halbert and Charles, listening to
the doleful dropping of the slow rain without, until the long-continued
suspense became intolerably painful. At length footsteps paused at the door;
there was a knock, and some one entered, and each drew a long breath as if
suddenly relieved, though Mrs. Melville started again, and became deadly
pale, when Charles Hamilton entered the room alone. He seemed much
agitated and distressed.
“Where is Halbert?” Mrs. Melville exclaimed; and her cry was echoed
by the others at the fireside. “Has anything happened to Halbert?”
“Nothing—nothing: Halbert is quite well,” said Charles, sitting down
and wiping the perspiration from his forehead, while Halbert’s wife clasped
her hands in thankfulness. “He will be here soon; but I come from a most
distressing scene—a deathbed—and that the deathbed of one who has spent
his life as an infidel.”
“A stranger, Charles?” asked Mary.
“A stranger, and yet no stranger to us,” was Charles’s answer; and he
pressed his hands on his eyes, as though to shut out the remembrance of
what he had so lately witnessed. As he spoke, the servants entered the room
for the usual evening worship, under the impression that the master had
returned; and Charles Hamilton took Halbert’s place; and wife, and
Christian, and the other Mary, marvelled when Charles’s voice arose in
prayer, at the earnest fervent tone of supplication with which he pleaded for
that dying stranger, that the sins of his bygone life might not be
remembered against him; and that the blood of atonement, shed for the
vilest, might cleanse and purify that polluted soul, even in the departing
hour; and to these listeners there seemed a something in Charles’s prayer, as
if the dying man and the sins of his fast fading life were thoroughly familiar
to him and them.

A dreary journey it was for Halbert and Charles Hamilton as they left the
warm social hearth and threaded the narrow streets in silence, following the
sick man’s messenger. It was a boisterous night, whose windy gusts whirled
the heavy clouds along in quick succession, scattering them across the dark
bosom of the sky, and anon embattling them in ponderous masses that
lowered in apparent wrath over the gloomy world below. A strange contrast
to the blithe house they had left was the clamour and rudeness of the
obscure inn they entered now, and an unwonted visitor was a clergyman
there; but up the narrow staircase were they led, and pausing for an instant
on the landing-place, they listened for a moment to the deep groans and
wild exclamations of impatient agony, as the sufferer tossed about on his
uneasy bed.
“Ay, sir,” said a servant, who came out of the room with a scared and
terrified expression upon her face, in answer to Halbert’s inquiry; “ay, sir,
he’s very bad; but the worst of it is not in his body, neither!” and she shook
her head mysteriously; “for sure he’s been a bad man, and he’s a deal on his
mind.”
She held open the door as she said so, and the visitors entered. The
scanty hangings of his bed hid them from the miserable man who lay
writhing and struggling there, and the brothers started in utter amazement as
they looked upon the wasted and dying occupant of that poor room; the
brilliant, the fashionable, the rich, the talented Forsyth—where were all
these vain distinctions now?—lay before them, labouring in the last great
conflict; poor, deserted, forlorn, and helpless, without a friend, without a
hope, with scarce sufficient wealth to buy the cold civility of the terrified
nurse who tended him with mercenary carelessness; pressing fast into the
wide gloom of eternity, without one feeble ray of life or hope to guide him
on that fearful passage, or assuage the burning misery of his soul ere it set
out. Halbert Melville, deceived by that poor sufferer of old, bent down his
face on his clasped hands, speechless, as the well-known name trembled on
his companion’s tongue,—
“Forsyth!”
“Who calls me?” said the dying man, raising himself fearfully on his
skeleton arm, and gazing with his fiery sunken eyes through the small
apartment. “Who spoke to me? Hence!” he exclaimed, wildly sitting up
erect and strong in delirious fury. “Hence, ye vile spirits! Do I not come to
your place of misery? Why will ye torment me before my time?”
His trembling attendant tried to calm him: “A minister,” she said, “had
come to see him.” He said: “He allow a minister to come and speak with
him?”
A wild laugh was the response. “To speak with me, me that am already
in torment! Well, let him come,” he said, sinking back with a half-idiotic
smile, “let him come”—— and he muttered the conclusion of the sentence
to himself.
“Will you come forward, sir?” said the nurse, respectfully addressing
Halbert. “He is composed now.”
Trembling with agitation, Halbert drew nearer the bedside, but when
those burning eyes, wandering hither and thither about the room, rested on
him, a maniac scream rang through the narrow walls, and the gaunt form sat
erect again for a moment, with its long arms lifted above its head, and then
fell back in a faint, and Halbert Melville hung over his ancient deceiver as
anxiously as though he had been, or deserved in all respects to be, his best
beloved; and when the miserable man awoke to consciousness again, the
first object his eye fell upon, was Halbert kneeling by his bedside, chafing
in his own the cold damp hand of Forsyth, with kindest pity pictured on his
face. Had Halbert disdained him, had he shunned or reproached him, poor
Forsyth, in the delirious strength of his disease, would have given him back
scorn for scorn, reproach for reproach. But, lo! the face of this man, whom
he had wounded so bitterly, was beaming on him now in compassion’s
gentlest guise; and the fierce despairing spirit melted like a child’s, and the
dying sinner wept.
“Keep back, Charles!” whispered Halbert, as he rose from the bedside;
“the sight of you might awaken darker feelings, and he seems subdued and
softened now. There may yet be hope.”
Hope!—the echo of that blessed word has surely reached the quick ear
of the sufferer; and it draws from him a painful moan and bitter repetition as
he turns his weary form on his couch again: “Hope! who speaks of hope to
me?”
“I do,” said Halbert Melville, mildly looking upon the ghastly face
whose eyes of supernatural brightness were again fixed upon him. “I do,
Forsyth; I, who have sinned as deeply, and in some degree after the same
fashion as you. I am commissioned to speak of hope to all—of hope, even
on the brink of the grave—of hope to the chief of sinners. Yes, I am sent to
speak of hope,” he continued, growing more and more fervent, while the
sick man’s fascinated attention and glowing eyes followed each word he
uttered and each motion of his lifted hand. “Yes, of hope a thousand times
higher in its faintest aspirations than the loftiest ambition of the world.”
“Ay, Melville,” he murmured, feebly overcome by his weakness and
emotion. “Ay, but not for me, not for one like me. Why do you come here to
mock me?” he added fiercely, after a momentary pause; “why do you come
here to insult me with your offers of hope? I am beyond its reach. Let me
alone; there is no hope, no help for me!” and again his voice sunk into
feebleness, as he murmured over and over these despairing words, like,
Charles Hamilton said afterwards, the prolonged wail of a lost soul.
“Listen to me, Forsyth,” said Halbert, seating himself by the bedside,
and bending over the sufferer. “Listen to me! You remember how I denied
my God and glorified in the denial when last I saw you. You remember how
I renounced my faith and hope,” and Halbert, pale with sudden recollection,
wiped the cold perspiration from his forehead. “You know, likewise, how I
left my home in despair—such despair as you experience now. Listen to
me, Forsyth, while I tell you how I regained hope.”
Forsyth groaned and hid his face in his hands, for Halbert had touched a
chord in his heart, and a flood of memories rushed back to daunt and
confound him, if that were possible, still more and more; and then, for there
seemed something in Halbert’s face that fascinated his burning eyes, he
turned round again to listen, while Halbert began the fearful story of his
own despair—terrible to hear of—terrible to tell; but, oh! how much more
terrible to remember, as what oneself has passed through. With increasing
earnestness as he went on, the poor sufferer gazed and listened, and at every
pause a low moan, wrung from his very soul, attested the fearful
faithfulness of the portraiture, true in its minutest points. It was a sore task
for Halbert Melville to live over again, even in remembrance, those awful
years, and exhibit the bygone fever of his life for the healing of that
wounded soul; but bravely did he do it, sparing not the pain of his own
shrinking recollection, but unfolding bit by bit the agonies of his then
hopelessness, so fearfully reproduced before him now in this trembling
spirit, till Charles, sitting unseen in a corner of the small apartment, felt a
thrill of awe creep over him, as he listened and trembled in very sympathy;
but when Halbert’s voice, full of saddest solemnity, began to soften as he
spoke of hope, of that hope that came upon his seared heart like the sweet
drops of April rain, reviving what was desolate, of hope whose every smile
was full of truthfulness, and certainty, firmer than the foundations of the
earth, more enduring than the blue sky or the starry worlds above, built
upon the divine righteousness of Him who died for sinners;—the heart of
the despairing man grew sick within him, as though the momentary gleam
which irradiated his hollow eye was too precious, too joyful, to abide with
him in his misery—and, lo! the hardened, obdurate, and unbelieving spirit
was struck with the rod of One mightier than Moses, and hiding his pale
face on his tear-wet pillow, the penitent man was ready to sob with the
Prophet, “Oh! that mine head were waters, and mine eyes a fountain of
tears!”
A solemn stillness fell upon that sick-room when Halbert’s eloquent tale
was told; a stillness that thrilled them as though it betokened the presence of
a visitor more powerful than they. The solitary light by the bedside fell
upon the recumbent figure, with its thin arms stretched upon the pillow, and
its white and ghastly face hidden thereon—full upon the clasped hands of
God’s generous servant, wrestling in silent supplication for that poor
helpless one. It was a solemn moment, and who may prophesy the issue, the
end of all this? A little period passed away, and the fever of the sick man’s
despair was assuaged, and weariness stole over his weak frame, with which
his fiery rage of mind had hitherto done battle; and gentle sleep, such as had
never refreshed his feeble body since he lay down on this bed, closed those
poor eyelids now. Pleasant to look upon was that wasted face, in
comparison with what it was when Halbert Melville saw its haggard
features first of all this night. God grant a blessed awakening.
Softly Halbert stole across the room, and bade Charles go; as soon as he
could leave Forsyth he promised that he would return home, but it might be
long ere he could do that, and he called the nurse, who was waiting without
the door, to see how her patient slept. She looked at him in amazement. Nor
was the wonder less of the doctor, who came almost immediately after—he
could not have deemed such a thing possible, and if it continued long, it yet
might save his life, spent and wasted as he was; but he must still be kept in
perfect quietness. Halbert took his station at the bedside as the doctor and
nurse left the room, and shading Forsyth’s face with the thin curtain, he
leant back, and gave himself up for a time to the strange whirl of excited
feeling which followed. The memories so long buried, so suddenly and
powerfully awakened; the image of this man, as he once was, and what he
was now. Compassion, interest, hope, all circled about that slumbering
figure, till Halbert’s anxiety found vent in its accustomed channel, prayer.
The night wore slowly on, hour after hour pealed from neighbouring clocks
till the chill grey dawn of morn crept into the sick-room, making the
solitary watcher shiver with its breath of piercing cold; and not until the
morning was advanced, till smoke floated over every roof, and the bustle of
daily life had begun once more, did the poor slumberer awake.
Wonderingly, as he opened his eyes, did he gaze on Halbert: wonderingly
and wistfully, as the events of the past night came up before him in
confused recollections, and he perceived that Halbert, who bent over him
with enquiries, had watched by his side all night. Forsyth shaded his eyes
with his thin hand, and murmured a half weeping acknowledgment of
thankfulness, “This from you, Melville, this from you!”
 CHAPTER III.

Hope the befriending,

Does what she can, for she points evermore up to heaven, and faithful
Plunges her anchor’s peak in the depths of the grave, and beneath it
Paints a more beautiful world * * * *
* * * Then praise we our Father in Heaven,
Him, who has given us more; for to us has Hope been illumined;
Groping no longer in night; she is Faith, she is living assurance;
Faith is enlighten’d hope; she is light, is the eye of affection;
Dreams of the longing interprets, and carves their visions in marble;
Faith is the sun of life; and her countenance shines like the Prophet’s,
For she has look’d upon God.—Evangeline.

HERE were anxious enquiries mingling with the glad welcome

which Halbert Melville received as he entered his own house on
that clear cold winter’s morning,—for the evening’s rain had
passed away, and frost had set in once more—enquiries that
showed the interest which both his own Mary and Christian—for
Christian’s society, though she did not allow it to be monopolised by either,
was claimed in part by both the Marys, and her time divided between them
—felt in the unhappy sufferer.
“Does Mary know, Christian?” was one of Halbert’s first questions.
“Yes,” was the answer, “and much was she shocked and grieved, of
course; as was Charles also, but we were all rejoiced to hear from him that a
happy influence seemed at work before he left you. Has it gone on? Can he
see any light yet, Halbert?”
“I dare not answer you, Christian,” said her brother gravely. “I know too
well the nature of Forsyth’s feelings to expect that he should speedily have
entire rest; but God has different ways of working with different
individuals, and I have reason to give Him thanks for my own terrible
experience, as I believe my account of it was the means of softening the
heart of yon poor despairing man.”
 “How wonderful, Halbert,” said Christian, laying her hand on his
shoulder; “how wonderful are the ways and workings of Providence. Who
could have imagined that you were to be the instrument, as I trust and pray
you may be, of turning your old tempter from the evil of his ways, and
leading him into the way of salvation!”
A month of the new year glided rapidly away, when one mild Sabbath
morning, a thin pale man, prematurely aged, entered Halbert Melville’s
church. The exertion of walking seemed very great and painful to him, and
he tottered, even though leaning on his staff, as he passed along to a seat. A
sickly hue was still upon his wasted features, and the hair that shaded his
high forehead was white, apparently more from sorrow than from years.
When he had seated himself, he cast around him a humble wistful glance, as
though he felt himself alone and begged for sympathy; and people of kindly
nature who took their places near him, felt themselves powerfully drawn to
the lonely stranger who looked so pale, and weak, and humble, and
wondered who he was; and many of them who watched him with
involuntary interest, noticed the quick flush that passed over his face as
Mary Hamilton entered, and how he gazed upon the other Mary, and
lingered with glistening eyes on every little one of the two smiling families,
as though their childish grace rejoiced his heart; but the observers wondered
still more when their minister had entered the pulpit to see the big round
tears which fell silently upon the stranger’s open Bible, and the expression
of almost womanly tenderness that shone in every line of his upturned face.
Mr. Melville, they said afterwards, was like a man inspired that day—so
clear, so full, so powerful was his sermon. His text was in one of Isaiah’s
sublime prophecies. “Look unto me and be ye saved, all the ends of the
earth, for I am God, and besides me there is none else.” And as he drew
with rapid pencil the glorious character of the divine speaker, in all the
majesty of the original Godhead, and also of his Mediatorial glory, his
hearers felt that he that day spoke like one inspired. Vividly he described
them lost in natural darkness, groping about the walls of their prison-house,
labouring to grasp the meteor light which flitted hither and thither about
each earthly boundary, hopeless and helpless, when this voice rang through
the gloom, “Look unto me and be ye saved.” Vividly he pictured the
entering light, which to the saved followed these words of mercy, steady,
unfailing, and eternal, that sprung from point to point of these desolate
spirit cells, illuminating the walls with heavenly radiance, and making them
prisons no longer, but changing them into temples dedicate to the worship
of the highest. “My brethren,” said the eloquent preacher, bending down in
his earnestness, as though he would speak to each individual ere he
concluded. “There are those among you who know the blessedness of being
thus plucked from the everlasting burnings—there are among you those
who have worn out years in a fiery struggle before they found rest;” and the
voice of the preacher trembled; “and there are those whose anguish has
been compressed into a little round of days; but I know also that there are
some here who can echo the words of one who knew in his own dread
experience the agony of despair:

“ ‘I was a stricken deer that left the herd

Long since, with many an arrow deep infixed
My panting sides were charged;’

and I rejoice to know that here there are those who can continue in the same
words—

“ ‘There ’twas I met One who had himself

Been hit by the archers, in his hands he bore
And in his pierced side, their cruel wounds;
With gentle force soliciting the darts,
He drew them out, and heal’d, and bade me live.’

and, oh, my brethren, did you but know the fearful suffering, the hopeless
anguish that follows a course of lost opportunities and despised mercies,
you would not need that I should bid you flee! escape for your lives, tarry
not in the cell, the plain fair and well watered, and like the garden of the
Lord though it seem; escape to the mountain lest ye be consumed. ‘Look
unto me and be ye saved, all the ends of the earth, for I am God, and
besides me there is none else.’ ”
The face of the lonely stranger is hidden, but those who sit near him are
turning round in wonder at the echoing sob which bore witness to the effect
of these thrilling words upon his mind; but when the minister had closed his
book, and the people united their voices in praise before the service ended,
the weak low accents of that humble man were heard mingling among
them, for he had found hope, even such hope and peace as the preacher of
this day had proclaimed in yonder dim sick-chamber to its dying occupant;
and this lowly man was he, raised as by a miracle at once from the gates of
hell, and from the brink of the grave. With gentle sympathy did Halbert
Melville, his work of mercy over, press the hand of that grateful man; with
kindly anticipation of his unexpressed wish did he bring the children one by
one before him, and they wondered in their happy youthfulness as the hand
of that slender stooping figure trembled on each graceful head; and when
the two little Marys hand in hand came smiling up Forsyth did not ask their
names. He discovered too clearly the resemblance shining in the daughter,
and scarce less distant in the niece of Mary Melville of old, and he
murmured blessings upon them. He feared to hear the name which brought
so many painful recollections in its sweet and pleasant sound.
But when a little time had passed away, Forsyth learned to love the very
shadow of Mary Melville’s eldest born, and cherished her as she sprang up
in graceful girlhood, as though she had been the child of his own old age,
the daughter of his heart. The solitary stranger was soon better known to the
hearers of the Rev. Halbert Melville, for he lingered about the place as
though its very stones were dear to him. Forsyth had made no friends in his
long season of sinful wealth and prosperity—gay acquaintances he had had
in plenty who joined his guiltiness, and called themselves friends, until the
new course of folly and excess on which he entered with headlong avidity
after Mary Melville rejected him, had dissipated his substance and made
him poor, and then the forlorn sufferer in his obscure apartment found out
the true value of these his heartless companions’ friendship. But now, a new
man among friends on whose unworldly sincerity he could rely without a
shadow of a doubt, his very worldly prospects brightened, and gathering the
remnants of his broken fortunes, he began now to use the remainder of
God’s once abundant gifts with a holy prudence, that made his small
substance more valuable a thousand fold, than the larger income that had
been so lavishly expended in the long years of his guilt and darkness; a
changed man was he in every particular, the talent which made him
foremost in the ranks of infidelity was laid upon God’s altar now, a
consecrated thing, and men who knew him first after his great changes,
marvelled at his strange humility, so unlike the world in its simple
lowliness. When he was told of the sinful and erring he bent his head and
blamed them not, for the remembrance of his own sins filled him with
gentlest charity, and when deed of mercy was to be done, that needed
earnest exertion and zealous heart, the mild and gentle Forsyth was ever
foremost delighting in the labour.
The threads of our tale have nearly run out; and we have but, as knitters
say, to take them up ere we finish. Our Halbert Melville is famed and
honoured; a wise and earnest minister, faithful and fervent in his pulpit,
unwearying in daily labour. His gentle Mary becomes the sweet dignity of
her matronhood well, rejoicing in the happy guardianship of these fair
children. Nor is the other Mary less blessed: the liberal heavens have rained
down gifts upon them all; seed-time and harvest, summer and winter, have
passed over their heads; but death and sorrow, making sad visits to many
homes around them, and leaving havoc and desolation in their train, have
never in their stern companionship come across these peaceful thresholds.
Now we must draw the veil, lest we should feel the hot breath of sickness in
these happy households, or see the approaching shadow of grief darkening
their pleasant doorways.
Our friend James grows rich apace; and were you to see his portly figure
and shining face “on Change,” where merchants most do congregate, you
would be at no loss to understand why his opinion is now so weighty and
influential. Messrs. Rutherford and Melville left a goodly beginning for
their more enterprising successor; and James is now a most prosperous,
because a most enterprising man. Robert, too, though at a distance in
another city, the resident partner of his brother’s great house, speeds well in
his vocation; and wedding one of his gentle sister-in-law’s kindred, has
made up our tale. The Melvilles are truly, as Mary said, a happy family.
But how shall we say farewell to our companion of so many days and
various vicissitudes—our generous single-minded Christian Melville; fain
would we linger over every incident of thy remaining story. Fain look upon
thee once more, dear Christian, in the sacred quietness of thine own
chamber, recalling the holy memories of the past. Fain go with thee through
thy round of duties, rejoicing in the love which meets thy gracious presence
everywhere. Fain would we add to our brief history another tale, recording
how the stubborn resolutions of a second Halbert would yield to no
persuasions less gentle than thine; and how the guileless hearts of the twain
Marys unfolded their most secret thinkings in sweetest confidence to only
thee; how thou wert cherished, and honoured, and beloved, dear Christian;
how willingly would we tell, how glad look forward through the dim future,
to prophesy thee years of happiness as bright and unclouded as this, and
testify to the truth of that old saying of Halbert’s, “that Christian would
never grow old.” But now we must bid thee farewell, knowing how “thy
soul, like a quiet palmer, travellest unto the land of heaven;” and believing
well that, Christian, whatever may happen to thee in thy forward journey,
however it may savour now, be it fresh trials or increased joys, will work
nothing but final good and pleasantness to thy subdued and heavenly spirit
—has not our Father said that all things shall work together for good to
them that love God as thou dost?—bringing but a more abundant entrance
at thy latter days into the high inheritance in thy Father’s Kingdom, which
waits for the ending of thy pilgrimage, dear Christian Melville.

THE END.

————
BILLING, PRINTER, GUILDFORD, SURREY.
 *** END OF THE PROJECT GUTENBERG EBOOK CHRISTIAN
MELVILLE ***

Updated editions will replace the previous one—the old editions will
be renamed.

Creating the works from print editions not protected by U.S.

copyright law means that no one owns a United States copyright in
these works, so the Foundation (and you!) can copy and distribute it
in the United States without permission and without paying copyright
royalties. Special rules, set forth in the General Terms of Use part of
this license, apply to copying and distributing Project Gutenberg™
electronic works to protect the PROJECT GUTENBERG™ concept
and trademark. Project Gutenberg is a registered trademark, and
may not be used if you charge for an eBook, except by following the
terms of the trademark license, including paying royalties for use of
the Project Gutenberg trademark. If you do not charge anything for
copies of this eBook, complying with the trademark license is very
easy. You may use this eBook for nearly any purpose such as
creation of derivative works, reports, performances and research.
Project Gutenberg eBooks may be modified and printed and given
away—you may do practically ANYTHING in the United States with
eBooks not protected by U.S. copyright law. Redistribution is subject
to the trademark license, especially commercial redistribution.

START: FULL LICENSE

Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade

Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.

Let us accompany you on the journey of exploring knowledge and

personal growth!

ebookfinal.com