- Sign-up Now!
 - Current Issue
 - Edit Your Profile/Unsubscribe

Subscribe | Media Kit | About Us | All Issues | Subscriber Feedback | Contact Us | Privacy Statement
Sunbelt W2Knews™ Electronic Newsletter
The secret of those "who always seem to know" - Over 500,000 Readers!
Sun, Nov 2, 1997
Sunbelt Windows NTools ENews[tm] Nov 2, 1997
  This issue of W2Knews™ contains:
Sunbelt Windows NTools[tm] Electronic Newsletter
Vol. 2, #21 - November 2,1997
--- http://www.sunbelt-software.com ---

***************************************************************
Sunbelt Windows NTools E-News is the world's first and largest 
E-Newsletter designed for NT System Managers that have the job of 
getting and keeping NT up & running in a production environment.
Sunbelt launched this electronic newsletter so that we could keep
members of the Windows NT community informed and aware of what is
happening with 3-rd party NT System Management Tools, and to
provide hints and tips that will enable you to better understand
and utilize Windows NT. You'll find general Windows NT related
and third party news, technical information, and 3-rd party beta
and release information. By subscribing to NTools E-News[tm], you 
are also a charter member of the Sunbelt Field Test Bonus Program.
Sunbelt Software is the first and largest distributor worldwide
of Third Party System Management Tools for Windows NT with
subsidiaries in the USA, UK, France, Germany, Holland and Italy.
****************************************************************

This issue of Windows NTools E-News contains:

1. "EDITORS CORNER" 
* MICROSOFT PREMIER SUPPORT INCREASES 30%-50% 
THE COST OF DOWNTIME

2. "TECH BRIEFING"
* YOUR SYSTEM THROUGHPUT AND NT MEMORY POOL FRAGMENTATION
* THE NT SERVER CRASHED: WHAT NOW?

3. "NT RELATED NEWS"
* TRIPP LITE LOOKS FOR BETA TESTERS 
* INTEL LANDESK SUPPORTS NT
* DIFFERENT NT MARKETS EMERGE

4. "THIRD PARTY NEWS"
* PERFMAN GETS BACKOFFICE CERTIFICATION
* FORTRESS-NT NOW ON ALPHA AND USED BY BOEING AND LOCKHEED
* ROLLING OUT OF MANY WORKSTATIONS: NEW SOLUTION
* NEW: BQMS for XLNT

5. "HINTS AND TIPS - TIME SAVERS AND OTHER GOODIES...

6. "NEW: THE NT STOCK WATCH"

7. "HOW TO USE THE MAILING LIST"
NEW Instructions on how to subscribe, sign off
and change addresses.

**************************************************************

1. "EDITORS CORNER" 

Hi All, 

As you may already have found out, Microsoft has increased
their Premiere Support Fees with about 30%-50%. (The reported
percentages are a bit different per country). From a user
perspective, this is not a particularly justifiable move.
Microsoft claims that this new fee will include a visit from
one of their engineers once a year but you can understand
that this has caused some people to wonder if it is worth
it.

Well, that depends completely on the cost of your downtime.
In the mini- and mainframe world this is the dreaded "d"-
word as in these environments you can easily get a cost of
$100,000 per hour of downtime. But this is getting more and 
more true of Mission Critial NT installations as well. If
you look at tens of thousands of NT Workstations on Wall 
Street you will see what I mean. One hour downtime of a top
trader can cost a few million.

Windows NT Servers do crash now and then. There are different
kinds of crashes as well. Sometimes it's a BSOD and a reboot
fixes the problem. Sometimes the server cannot be booted anymore
and then the real trouble starts. A recent survey we did showed 
this happens on average once every 9-16 months and workstations 
crash (unbootable) more often. The latter probably due to user 
errors one would presume. ;-)

But what IS the cost of unplanned downtime? For some reason
the system crashed and you are in an sudden "unscheduled
maintenance" crisis. A server that is operational 99% of the
time is still going to conk out on you 80 hours per year...

Calculating the cost is a total of many different elements:
hours from users, lost opportunities, perhaps lost data that
needs to be rekeyed or backed up etc etc. The downtime cost
will vary greatly in different organizations, but with 50 to 
100 users on a server, you can easily count with at least a 
thousand bucks per hour, very likely double that.

So if you pay Microsoft let's say $20,000 for Premiere support
per year, is it worth it? Well, if with their help your
downtime goes from 99% to 99.95% (from 80 down to 4 hours)
it's obviously worth it. But that is hardly realistic, there
will often be a bunch of calls up and down to get the problem
resolved.

Microsoft is in a position where it can try to raise prices 
and see if they can get away with it. Usually companies find
out when customers vote with their feet. Don't buy support
and subscribe to Technet ($300/yr.) Read the industry mags
like Windows NT Magazine, BackOffice, NT Systems and subscribe
to list servers like NTSYSADMIN where you can discuss your
problems pretty much in real time with thousands of colleagues.

And when you _have_ a really nasty problem, call Microsoft on 
a call-by-call basis for about $200 a pop. 70% of NT servers
are installed in outfits with 100 staff or less. Premiere 
support is only worth it for very large outfits or consulting
firms, and with prices going up like this, it makes sense to
grow in-house MCSE's that will pick up those problems. MCSE's
get paid on average a minimum of $5,000 per year more than 
uncertified co-workers. Is that 5 grand going to prevent more
than 2 or 3 hours of downtime? You Betcha!! Becoming an MCSE
also get 60% of them moved to a better position within a year.

An additional suggestion is to have a tool around to get
unbootable servers or workstations back up. This saves tons
of time. More on this in the Tech Briefing.

In other words, get your NT Support in your own hands if you
possibly can, it will save a LOT of money with the prices
going up like this!

Warm regards,

Stu Sjouwerman


************************************************************** 
2. "TECH BRIEFING"

* YOUR SYSTEM THROUGHPUT AND NT MEMORY POOL FRAGMENTATION

"WHAT? WHAT THE HECK IS POOL FRAGMENTATION?" You were probably 
wondering when you read that header. Well, let's start at
the beginning and look at throughput first. In the PC space
performance is generally measured in CPU speed, disk access, 
memory nanoseconds and the like. But in the real NT server 
environment, those are only a secondary type of consideration.

The real benchmark is System Throughput. A simple definition
is "The volume of work done in a fixed amount of time". This
is often measured in "Transactions per Second" or Read/Write
Disk I/O's per second, or # of Memory Read/Writes per Second.

If you run a certain benchmark, make one change, run it again
and look at the "before-and-after" numbers you have a throughput
measurement. That is a much more honest way to determine perfor-
mance of a server. The whole system is looked at, the combination
of hardware, O/S, applications and network access speeds.
WinBench is a very well known benchmark tool that shows real
life throughput metrics and allows you to identify performance
enhancements.

So, what has throughput to do with fragmentation of the memory
pools you might ask?

Let's have a quick look at NT's memory management scheme, but 
first some definitions:

- Physical Memory: The actual amount of RAM in your system
- Virtual Memory: The amount of RAM each application 'perceives'.
(BTW this is 4Gig split in half, 2Gig for the app, 2Gig for 
the system areas)
- Paging file: data that is not immediately necessary in RAM gets
flushed to this file on the hard drive.
- Pagefault: the actual process of writing data from RAM to the
pagefile on disk
- Working Set: the data in RAM a program works with at a certain
moment, this is dynamic, it shrinks and expands based on system
resources
- NonPaged Pool: Not all data can be paged out to disk. Some of it
needs to stay in RAM and cannot be paged, hence the nonpaged
memory pool
- Paged Pool: Obviously the data that _can_ be written to disk if
needed.

With these definitions things should start to become clear. Each
app has a virtual space of 4Gig. But NT loads a minimum of data
in RAM. The rest can easily be accessed from the page file if 
needed. When a page of data is required, it gets read from the 
page file into RAM, used and if not used for a while, faulted back
to the page file. As you can see, NT memory management is basically
a simple concept, the disk is used as an extension of RAM, and data
gets faulted in and out of RAM as needed. 

Now, what is FRAGMENTATION of these memory pools? 

Blocks of data get written to and deleted from these pools at high 
speeds. Not all blocks are the same size. This high speed process 
of reads and deletes fragments the pools, you could compare it to 
disk fragmentation, but it is not the same.

If your pools are full of memory blocks that are no longer in use,
or were allocated by terminated applications, or services that are
no longer running, the process of memory allocation starts to become 
less efficient. If you could free up all the unused blocks and move 
all the used blocks to the beginning of the pool, the result would be
a more contiguous pool with faster access and getting rid of the
"holes" in the pool.

And THAT my friends, will speed up your throughput! 
-----------------------------

* THE NT SERVER CRASHED: WHAT NOW?

OK, the BSOD stuck up its ugly blue head and the server is unbootable.
Shit happens: data corruption in one of the thousands of NT DLL's,
improperly installed or ill behaved apps, faulty configuration,
bad sectors on the O/S disk, device drivers that do not play well
with the kernel, you name it.

There are several options to handle an emergency like this:


1) Cheyenne has disaster recovery tools - 

Cheyenne's tools are backup solutions that can create a boot floppy
and restore the system from tape. It is time consuming and certainly
not the ideal way to get back on line fast but it works and is better
than nothing.

2) Boot Sector Repair

This handles a relatively rare class of problems. Much more common is a
corrupt or missing file, a Registry entry that is incorrect, or the
addition of a new driver or piece of software that prevents the machine
from booting successfully. 

3) Repair Disk Utility - (the native NT ERDISK)

Its only as good as how current the disks are kept, and it many times 
fails to detect corrupt drivers and update them with fresh versions. 
Its file and registry fixes are whole-sale. It would be better to 
do precision fixes.

4) Doing a parallel install of NT to another partition, manually 
rebuild system, restore from tape.

This is definitely an option, but many times there is no room on the 
system to do a parallel install, and doing so takes at least 1/2 hour. 
Then the complete backup from tape on top, and your users sitting on
their hands. Takes even longer than the Cheyenne solution.

Conclusion: Ideally you should have a utility that allows you to
access the dead server from another system. You could boot the dead
server with a floppy and create a fast serial connection. No space
needed on the disk of the dead server and immediate access to the
data for a pinpoint surgical repair. In the mean time, better
religiously make those ERdisks and hope for the best.
--------------------------

**************************************************************
3. "NT RELATED NEWS"

* TRIPP LITE LOOKS FOR BETA TESTERS

Tripp Lite is one of the largest UPS systems manufacturers and 
develop software for the NT market. They have some exciting 
software in the pipeline, but are looking for beta testers.
UPS software has always been able to shutdown the OS, but 
applications and open files were left flapping in the virtual 
wind. They have some new AI technology that will discover all open 
apps, files, etc. and save the active data before the OS shutdown. 
With half of the installed NT servers installed as app servers, 
the success of thin client/server products like Citrix, and the 
eventual multi-user version of NT, Tripp Lite is completely 
focused on protecting the data at the server. They have tested 
with Oracle 7.3, Notes, Domino Server Office versions,etc. So far 
we have a 90% success rate with the AI closing and saving things, 
but they are seeking other apps to test against. Feel like some
good 'ol crashin servers? Send an email to T Chin at Tripplite
at tchin%[email protected] and tell him you're interested.
-------------------------------

* INTEL LANDESK SUPPORTS NT

Intel last week unveiled LANDesk Management Suite 6, which adds support 
for Windows NT. The new version will let network managers control NT 
and IntranetWare servers from a single point. The upgrade provides 
software distribution, remote control, hardware and software inventory, 
software metering, virus protection, and server management. It also 
supports any ODBC- compliant database. That NT support was High Time!
--------------------------------

* DIFFERENT NT MARKETS EMERGE

It's too easy to pronounce markets as similar to one another if they 
use the same kinds of product. However, The fact is, that belief of 
product-centric markets is no longer true. Even in a market as specicially
as Windows NT, because different players have unique selling, marketing, 
and support strengths, they're showing up in different segments.

For example, if we look at the size of companies using NT, you can see
that there are significant differences in the vendors who are successful 
in smaller companies and those that do better in bigger firms. For
instance Compaq, HP, and IBM are much more predominant in larger firms. 

But others are very good in small business and obviously it is all 
sales and marketing based, not product based. A good example are 
Dell and Gateway with direct selling models. Their ability to use the 
lower cost sales method of outbound telesales puts them in more 
accounts, and as a result that gives them better results.

The NT market is not just fragmented by size of firm, but there is 
also an increasing level of "vertical-ization", or development of 
specific industry solutions, going on with the NT market. Looking 
at major industry groups, you can see that competitors are focused on 
different segments.

For example, in the very important services industry, Dell and Gateway
have larger share, followed by HP. It's also worth noting that the 
servicesindustry group is comprised of smaller firms. On the other 
hand, the finance-insurance-real estate (F-I-R-E) industry are more 
common in larger firms, and IBM and Compaq are the primary players 
here. Gateway and IBM are the leaders in wholesale/retail. So you 
start to get the picture that there's a very different competitive 
set based upon the segments in which different vendors have success.

The NT market is a huge opportunity in general, but really, it is made 
up of a lot of smaller opportunities united by a common software 
platform and a very diverse group of hardware vendors.
-----------------------------


**************************************************************
4. "THIRD PARTY NEWS"

* PERFMAN GETS BACKOFFICE CERTIFICATION

PerfMan, the early warning system for NT Lans, now proudly
carries the Microsoft Backoffice Logo. The logo indicates
that the product has been torture tested by the Veritest
Labs and is declared compatible with the BackOffice Family
of products. 

Perfman is a very interesting tool. It does not use agents
that need to be installed on each managed system, the data
is pulled _off_ those boxes by the central PerfMan server.
It is also incredibly efficient with CPU. You can manage
250 machines easily from one server, so it scales to a multi
thousand workstations and server environment. 

Windows NT Magazine LAB Reports wrote this about it:
" If you use PerfMan consistently and interpret the gathered
data regularly, PerfMan will help you which areas of your
network need improvements, gauge which servers need hardware
upgrades, learn which machines are overused, ascertain which 
systems are improperly located in the logical layout, and
prevent system failures and performance dips.

"By identifying areas where improvements will have the greatest
benefit, you can also use the data that PerfMan has gathered
to plan future network growth and expansion. For example, you
can measure the throughput, CPU usage, and disk access to see
where you need network upgrades, CPU replacements, and larger
or faster hard disks"

It's a complete trending, capacity planning and management 
reporting tool that can be set up in a fully automated way.
Set it up and walk away! 

Ask yourself: Do I Manage By Crisis Or Prediction? PerfMan is 
a super value compared to some other performance solutions, 
and takes you about an hour a week to check to see how things
are going. A great time saver and excellent downtime prevention
utility.

The best way to find out about PerfMan is to try it out for 
yourself, download the latest version from: 
http://www.sunbelt-software.com/perfman.htm
--------------------------------

* FORTRESS-NT NOW ON ALPHA AND USED BY BOEING AND LOCKHEED

Fortress-NT has been ported to the Alpha platform and is getting
very popular in secure sites. It is a tamper proof screen saver
and idle process killer. You can force a logoff after any given
time period and also at a certain time in the evening to make
SURE your backups will not skip any file left open by a user.

This is one of those really useful simple tools that make sure
your security is forced in, your backups have the highest chance
of success and also help performance to stay up. Fortress-NT was
made especially for WinNT workstation but will also run on SV.

If you live in a secure MIS environment like the government, or
bank, or insurance company this is actually something you should
look at, it can easily be remotely installed. The US Coast Guard 
uses Fortress-NT as well. Features the "FortLock" function that
lets users quickly lock their workstation with a simple button
click, taskbar icon or hotkey press.

Fortress-NT helps to keep C-2 security standards applied. Download 
an eval copy at: http://www.sunbelt-software.com/fortress.htm
-----------------------------

* ROLLING OUT OF MANY WORKSTATIONS: NEW SOLUTION

I have run across a new utility for those rolling out new systems or
upgrading existing ones. Unlike Ghost, Imagecast broadcasts the system
image to multiple CPUs at one. They call this Multicasting. Test
it yourself and see if this type of utility will ease the pain of
multi-system roll outs. NetVersant also claims to have a method to
handle the SID problem which will be released by Comdex Fall this month.
Be aware of support issues related to this SID matter as discussed
in earlier issues. Microsoft should _really_ understand the viewpoint of
people that need to roll out thousands of systems and help them to
do this smoothly. The people at NetVersant are very responsive to your 
needs though, You can check ImageCast out at the NetVersant web site: 
http://www.netversant.com
------------------------------

* ENHANCED VERSION 4.0 OF PROXY PLUS SERVER

CSM announced the 4.0 version of their acclaimed CSM Proxy Plus
security and access control server that doesn't require an in-
house networking guru to administer or configure. The new and 
substantially enhanced CSM Proxy Plus server performs as a gateway 
proxy, caching, virtual private network server and firewall system. 

With the increasing need to connect a corporate LAN to the Internet 
with security and access controls like Who, What, When and Where; 
an inexpensive but robust solution like Proxy Plus is the ideal 
solution. "CSM Proxy Plus is an 'All-In-One' Internet access 
utility," explains Mark Smith, President, CSM-USA. "It serves as a 
gateway for the LAN to access the Internet through a single IP 
address/connection, provides caching for bandwidth savings, and 
controls access and intrusion. In addition, companies use CSM 
Proxy Plus to establish a secure Virtual Private Network.

"We have been very impressed with the ease-of-use and stability 
displayed by CSM Proxy Plus," states Carl Schmidt, President of 
Applied Information Technologies. "We've used it in installations 
ranging from dial-up 28.8KBps to dedicated 10MBps, and it's worked 
flawlessly with no configuration hassles. This is an essential tool 
for every MIS manager's Internet toolkit. We're amazed at the wealth 
of features offered by CSM Proxy Plus".

CSM Proxy Plus 4.0 supports all major internet protocols, and can
be checked out at: http://www.csm-usa.com
-----------------------------

* NEW: BQMS for XLNT

The XLNT Batch Queue Management System (BQMS) Technology Preview will
showcase beta software that is scheduled for inclusion in XLNT V2.
BQMS is a completely distributed batch queue and scheduling system. 
BQMS provides support which allows workstations (clients) to submit 
batch jobs (running the XLNT command language or anyother command 
shell) to one or more queue management machines (servers). The queue 
managers provide the ability to create execution and generic batch 
queues. An execution batch queue is associated with a machine and 
allows job submitted to that queue to be run on the named machine. 
An execution queue can also run on the same machine as the BQMS 
queue manager. A Generic queue consists of one or more execution 
queues and provides the user with the ability to load share and/or 
failover execution of jobs within the queue system.

BQMS includes both Window GUI and XLNT style command line support. 
Among the BQMS Queue commands supported are INITIALIZE, MODIFY, 
PAUSE/RELEASE, DISPLAY, START/STOP, OPEN/CLOSE, FLUSH, REDIRECT, etc. 
BQMS Job commands include SUBMIT, HOLD/RELEASE, DELETE, DISPLAY and 
SYNCHRONIZE. 

XLNT provides language facilities which further assist batch job
scripting. The F$MODE lexical determines how the job was initiated 
(for example, interactive or batch). SET RESTART_VALUE allows the 
script writer to employ the built-in batch job checkpoint facility of 
BQMS. The restart value is typically the last label or section within 
the script that represents the current job step. If the batch job is 
restarted (a BQMS Job attribute), the $RESTART symbol can be tested 
(to determine if the batch job has been restarted) and the 
BATCH$RESTART symbol is provided with the value that you specified 
as part of the SET RESTART_VALUE command.

BQMS also provides notification support, output logging of the batch 
job itself and because BQMS is part of XLNT, batch script writers enjoy
complete error and exception handling that's a major requirement of 
serious batch or non-interactive process handling.

This technology preview will be initially available for the Intel 
Windows NT platform only. You must be running XLNT V1.1 SP2 (or later) 
to successfully try the BQMS preview. Check it out at:
http://www.advsyscon.com
------------------------------


**************************************************************
5. "HINTS AND TIPS - TIME SAVERS AND OTHER GOODIES...

The latest jargon watch:

- Meatspace: the space that begins where fingertips reach
the keyboard. As opposed to Cyberspace that begins at
the keyboard and into the net. Used to be called "Reality".

- Billy Backhoe: Anyone that damages network or transmission
lines.

- Cata: Network slang for catastrophic failure, something
network managers do not like to say.

- Code 13: A problem with no clear cause, such as an inter-
mittant failure.

- Smoking Gun: A clear cause of a problem, such as a carrier
failure or a user installing a ill behaved Win95 game on 
an NTWS.
---------------------------

BEFORE you upgrade 3.51 to 4.0, remove or disable any non-
Microsoft device drivers or services to minimize migration
problems.
---------------------------

CPU pegged out most of the time, and in need of a second one?
NT may not realize you just plugged in a second CPU, in the
NT4.0 resource kit is a utility called UPTOMP.EXE you need
to run. That will make NT wake up and smell the java.
---------------------------

FAT or NTFS? If you need to make a choice, only use FAT for
compatibility issues on NTWS. A simple rule of thumb is that
any partition over 400meg will run better and more reliable 
with NTFS.
---------------------------
**************************************************************
6. "NEW: THE NT STOCK WATCH"

52 WK 52 WK P/E WEEK
SECURITY CLOSE HIGH LOW RATIO CHNG
---------------------------------------------------------------------
Dell Computer Corp....... 80 1/8 103 7/8 19 7/32 40 -14.7%
Digital Eqpmt. Corp...... 50 1/16 52 5/16 25 39 -0.9%
Intergraph Corporation... 10 3/4 14 3/16 6 1/4 -9.4%
McAfee Associates Inc.... 49 3/4 78 1/2 36 1/2 32 -3.6%
Microsoft Corporation.... 130 150 3/4 68 1/4 49 -3.9%
Network General Corp..... 20 1/4 30 1/4 11 1/8 -3.5%
Novell Inc............... 8 7/16 13 6 17/64 -6.2%
Oracle Corporation....... 35 49/64 42 1/8 22 25/64 50 +5.4%
Qualcomm Inc............. 56 3/8 68 1/2 38 58 -4.8%
Qualix Group Inc......... 4 7/16 9 3/4 4 14 -4.7%
Seagate Technologies Inc. 27 1/8 56 1/4 25 1/8 23 -10.6%
Sun Microsystems Inc..... 34 1/4 53 5/16 25 1/2 18 -8.8%
Sybase Inc............... 16 5/16 23 5/8 12 1/8 65 -11.2%
Symantec Corporation..... 21 7/8 25 3/4 9 7/8 30 +4.7%
Dow Jones 30 Industrials. 7,442.08 -3.5%


**************************************************************
7. "HOW TO USE THE MAILING LIST"
Instructions on how to subscribe, sign off
and change addresses

TO SUBSCRIBE TO THE LIST

Go to: 
http://www.sunbelt-software.com/scripts/lyris.pl?join=nt-list
and fill out the form, simple & easy: 1 minute work.
_____________________________________________________

TO QUIT THE LIST

Two ways to do it: 

1) Go to:
http://www.sunbelt-software.com/scripts/lyris.pl 
choose the NT-List, use your email address that is at the
bottom of each newsletter and leave the list via the
web interface.

2) Simply forward any newsletter you get to this email address:
[email protected]
_____________________________________________________

TO CHANGE YOUR ADDRESS

First unsubscribe and then resubscribe as per the
procedure above.

*************************************************

FOR MORE INFORMATION

On the World Wide Web point your browser to:

For the newsletter:
http://www.ntnews.com
For our website:
USA: http://www.sunbelt-software.com

Email for US sales information to:
[email protected]
Email for US Tech support to:
[email protected]
Email to the US Editor:
[email protected]

Email for European Sales to:
[email protected]
Email for European Tech support to:
[email protected]

**************************************************

Legal Stuff:
This document is provided for informational purposes only. 
The information contained in this document represents the
current view of Sunbelt Software Distribution on the issues
discussed as of the date of publication. Because Sunbelt
must respond to changes in market conditions, it should not
be interpreted to be a commitment on the part of Sunbelt
and Sunbelt cannot guarantee the accuracy of any
information presented after the date of publication.

INFORMATION PROVIDED IN THIS DOCUMENT IS PROVIDED "AS IS"
WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESS OR IMPLIED,
INCLUDING BUT NOT LIMITED TO THE IMPLIED WARRANTIES OF
MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
FREEDOM FROM INFRINGEMENT.

The user assumes the entire risk as to the accuracy and the
use of this document. This document may be copied and
distributed subject to the following conditions: 1) All text
must be copied without modification and all pages must be
included; 2) All copies must contain Sunbelt's copyright
notice and any other notices provided therein; and 3) This
document may not be distributed for profit. All trademarks
acknowledged. Copyright Sunbelt Software Distribution,Inc 1997.

(email me with feedback: [email protected])