home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Telecom
/
1996-04-telecom-walnutcreek.iso
/
book.reviews
/
vaxcluster.principles
< prev
next >
Wrap
Text File
|
1995-01-01
|
4KB
|
80 lines
Date: 16 Oct 93 13:21 -0600
From: Rob Slade <roberts@decus.arc.ab.ca>
Subject: Book Review: "VAXcluster Principles" by Davis
BKVXCLST.RVW 930910
Digital Press
PO Box 3027
One Burlington Woods Drive
Burlington, MA 01803-9593
800-DIGITAL (800-344-4825)
VAXcluster Principles, Davis, EY-M740E-DP-CRE, 1993
If "principles" were all this book dealt with, it would only be
fifteen pages long. What we have here, rather, is a very detailed
description of the component parts of a VAX cluster - other than the
VAXen themselves. Communications channels and devices, storage
hardware, and the various system files and utilities are examined.
Included in this overview are some of the underlying principles of
those components.
Although it is somewhat unfair to ask a technical work to conform to
the same literary standard as other works, the style here could
definitely use improvement. We are given the same (word for word)
definition of a VAX cluster twice in the first eight pages, and the
same (Digital specific) definition of a network twice within twelve
pages. I know we are talking about fault tolerance here, but some of
the redundant material could easily be removed. At the same time, the
formatting of the text and the use of bullets and indenting is often
confusing, especially when multiple levels of bullets and indents get
mixed on a single page.
Since the VAX cluster is implemented only on proprietary machinery and
a proprietary operating system, it is not surprising that much of the
material is hardware and operating system specific. Scattered
throughout the work, however, are gems of generic interest, such as
the well-defined structure of an Ethernet packet, right down to the
diagrammed description of Manchester phase encoding of data.
Chapter one is an introduction to the topic. Unfortunately, it is
possibly the most confusing part of the book. Chapter two, discussing
the outline of Systems Communication Architecture (SCA) is somewhat
better, but still a bit ragged. (For example, having almost completed
the description of SCA, we are given a simplistic analogy of SCA
likened to a telephone system. This might better have been an
introduction.)
Davis is obviously more at home with the guts of the system, and the
book improves a great deal as chapters three through eight delve into
interconnects, storage architecture and options, locking, the
Connection Manager, and miscellaneous topics.
The work is replete with specific examples and notes regarding
particular devices. It is not simply a reproduction of the
documentation and marketing notes, though. The material is
well-organized, and gives details about the underlying structures
before moving on to details of operation. This may be frustrating for
some who merely want a "cookbook", but for the curious it can be an
interesting and potentially valuable excursion.
For those considering the installation or configuration of a cluster
the book will give valuable background and guidance. Those already
managing VAX clusters may not find an immediate need for it, but will
probably deepen their understanding of the system.
Those running MVS, of course, won't have the slightest idea what I'm
talking about.
copyright Robert M. Slade, 1993 BKVXCLST.RVW 930910
Permission granted to distribute with unedited copies of the TELECOM
Digest and associated newsgroups/mailing lists.
Vancouver Institute for Research into User Security, Canada V7K 2G6
Robert_Slade@sfu.ca ROBERTS@decus.ca rslade@cue.bc.ca Fidonet 1:153/733
p1@CyberStore.ca 604-526-3676