HP D7171A - NetServer - LPr Instruction Manual

HP NetServer

Performance and Sizing

Guidelines

Microsoft Windows NT 4.0 Server,

Terminal Server Edition

January 1999

Hewlett-Packard Company 2

Notice

The information contained in this document is subject to change without notice.

Hewlett-Packard makes no warranty of any kind with regard to this material, including, but not

limited to, the implied warranties of merchantability and fitness for a particular purpose.

Hewlett-Packard shall not be liable for errors contained herein or for incidental or consequential

damages in connection with the furnishing, performance, or use of this material.

The use of trademarks or other designations is for reference purposes only. The following

trademarks are used in this manual:

Intel, Pentium and Xeon are registered trademarks of Intel Corporation. Microsoft and MS-

DOS are U.S. registered trademarks of Microsoft Corporation. Windows and Windows NT are

trademarks of Microsoft Corporation. Citrix WinFrame and ICA are registered trademarks of

Citrix Systems, Inc.

Hewlett-Packard

Network Server Division

10955 Tantau Avenue

Cupertino, California 95014

Hewlett-Packard Company 3

Contents

2 TEST SUMMARY............................................................................4

Introduction......................................................................................................................4

Processors.........................................................................................................................5

Memory............................................................................................................................7

Network............................................................................................................................7

Client Solutions ................................................................................................................8

3 SIZING ACTIVITY..........................................................................10

Methodology...................................................................................................................10

Test Configuration .......................................................................................................10

Visual Test Scripts.......................................................................................................14

Test Procedures............................................................................................................16

Test Results ....................................................................................................................16

Elapsed Time...............................................................................................................17

Process Utilization........................................................................................................18

Network Utilization......................................................................................................24

Memory Utilization......................................................................................................26

Hewlett-Packard Company 4

2 Test Summary

Introduction

In a Microsoft® Windows NT® 4.0, Terminal Server (TSE) multiuser

environment users share hardware and software resources on the server.

These resources include processors, memory, and storage as well as operating

system resources and application resources such as configuration data (e.g.,

location of directories and files, arrangement of application within groups, and

desktop appearance).

In order to help standardize sizing information, Microsoft has defined three

categories of users:

• Task-based users. A task-based user typically runs a single application

used for data entry, for example a line-of-business application written in

Visual Basic (POS, Order Entry applications, SAP, and so on).

• Typical user. A typical user runs one or two applications but usually runs

only one at a time. A typical user runs applications whose data processing

demands on the system are not heavy, for example a productivity

application such as a word processor, a web browser or an e-mail client.

• Advanced users. An advanced user is a more sophisticated user who runs

three or more applications, and often has several applications active at the

same time. An advanced user runs applications whose data processing

demands on the system are heavy.

For the sample configurations described in the following table, Microsoft

provides, for each type of user, the recommended number of users that can be

supported (Microsoft Window NT Server, Terminal Server Administrator's

Guide).

Configuration Task-based users Typical users Advanced users

Single processor Intel®

Pentium®Pro 200MHz

128MB RAM

25 15 8

Dual-processor Pentium Pro

200MHz 256MB RAM 50 30 15

Quad-processor Pentium

Pro 200MHz 512MB RAM 100 60 30

Table 1

Hewlett-Packard Company 5

Processors

In the TSE environment, all users contend for the same processors. In this

environment a single CPU-intensive application can adversely effect the

performance of all users on the system.

As expected, our results also show that as the CPUs become saturated,

performance deteriorates significantly and client sessions can become starved.

As a result, solutions consisting of well behaved, predictable applications (e.g.,

order entry applications) are ideally suited for this environment.

The following table contains general guidelines, for the number of users that

can be supported by on a sample set of HP NetServer configurations.

HP NetServer

Configuration Task-based users Typical users Advanced users

LX Pro: Quad processor

Pentium Pro 200 MHz

768MB RAM

100 60 30

LPr: Dual processor

Pentium II 450 MHz 768MB

RAM

110 65 32

LH 3: Dual processor

Pentium II 450 MHz 768MB

RAM

110 65 32

LXr 8000: Quad Processor

Pentium II Xeon™400

MHz 1GB RAM

200 120 60

Table 2

Resource requirements can vary significantly depending on the application mix

and client usage pattern. Therefore, whenever possible, HP recommends

sizing the server using the targeted application environment.

Hewlett-Packard Company 6

Our tests also show that TSE scales well, evenly sharing the processors

among the client sessions.

The following chart shows the average and maximum processor utilization for

each of the four processors during the 1, 12, 24, 36, 48, 60 and 72 client runs

on the HP NetServer LXr 8000. This chart illustrates how evenly processing is

shared among the processors.

% Processor Utilization

100

Number of Client Sessions

Processor 0(Avg) 0.01 0. 05 0. 3 1.66 5. 19 11.14 25.01 37.64

Processor 1(Avg) 0.04 0.08 0.37 2.02 4.77 11.81 25.26 37.66

Processor 2(Avg) 0.01 0.04 0.45 1.68 5.18 11.34 25.52 38.05

Processor 3(Avg) 0.1 0. 03 0.59 1.81 5.13 11.63 25.35 37.79

Processor 0(Max) 0.21 0.78 3.59 6.56 27.99 61.94 100 100

Processor 1(Max) 1.2 1.46 4.35 11. 03 27.34 62.4 100 100

Processor 2(Max) 0.42 0.47 5.47 7.15 28.41 63.85 100 100

Processor 3(Max) 1.09 0.52 5.93 7.54 28.12 63.34 100 100

0 1 12 24 36 48 60 72

Figure 1

Hewlett-Packard Company 7

Memory

As with processors, configuring the server with insufficient memory will

seriously impact performance. Since memory requirements will vary, it is

extremely important to accurately size your server configuration using the

targeted application environment.

For Terminal Server, Microsoft requires 32MB of RAM, plus an additional 4 to

8MB RAM for each user to support running the user's applications on the

server (Microsoft Window NT Server, Terminal Server Administrator's Guide).

Memory utilization during our benchmark test averaged between 5 and 9MB

per user.

Network

As with processor and memory utilization, network utilization will vary

depending on user activity. As a general rule of thumb, Microsoft recommends

2 - 6Kbps per user (Microsoft Terminal Server Capacity Planning, White

Paper).

Network activity during our benchmark test averaged between 2 and 5Kbps per

user.

Hewlett-Packard Company 8

Client Solutions

Microsoft Terminal Server supports a full range of client choices including

Windows-based Terminals (WBTs), Palmtop PCs, NetPCs, and PCs. This

flexibility allows customers to select the client technology that best meets their

needs. HP, as a leading provider of both client and server technology, is

uniquely positioned to offer a full range of client/server solutions for the

Microsoft Terminal Server market.

Our tests have shown that the type of client (i.e., WBT, NetPC, or PC) is not a

significant factor when sizing the server. In order to illustrate this result the

same test script was run on both a PC configuration and a WBT configuration.

NetSe rverLHx

10 Mbits

Microsoft Terminal Server

HP NetServer

Microsoft Terminal Server Clients

HP Vectra

HP J31218A

PC Configuration

NetServerLHx

10 Mbits

Microsoft Terminal Server

HP NetServer

Microsoft Terminal Server Clients

HP Windows-based Terminals

HP J31218A

WBT Configuration

The next three charts illustrate that server behavior (i.e., performance and

resources utilization) does not vary significantly with a change of client

technology.

The following chart shows the average elapsed time recorded during 1, 2 and 3

client runs using the HP Vectras and HP Windows-based Terminals.

Average Elapsed Time

2000

2100

2200

2300

2400

2500

2600

2700

2800

2900

3000

Number of Client Sessions

PCClients 2440 2435 2431

WBTClients 2444 2438 2434

123

Figure 2

Hewlett-Packard Company 9

The following chart shows the average processor utilization recorded during 1,

2 and 3 client runs using the HP Vectras and HP Windows-based Terminals.

Average Processor Utilization

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Numberof Client Sessions

PCClient s 0.12 0.23 0.34

WBTCl ient s 0.13 0.23 0.35

123

Figure 3

The following chart shows the % Network Utilization recorded during 1, 2 and 3

client runs using HP Vectra and HP Windows-based Terminals.

Network Segment - Average % Network

0.5

1.5

2.5

umber of Client

PC 0.1

.26

.37

WBT 0.1

.26

.37

Figure 4

Hewlett-Packard Company 10

3 Sizing Activity

Methodology

This sizing activity consisted of running the same test script on various HP

NetServer configurations and then comparing the results.

A Visual Test 4.0 script, based on the Microsoft Office 95 product suite, was

used to automate and standardize the interaction of users (i.e., clients) with

Microsoft Windows NT 4.0 Server, Terminal Server Edition using the Microsoft

RDP protocol. (Visual Test is a product of Rational Software,

www.rational.com.)

For each TSE client session, the time to execute the test script (i.e., the

elapsed time) was recorded. In addition to elapsed time, additional system

utilization statistics were gathered using the Microsoft Performance Monitor.

Test Configuration

The test configuration consisted of an HP NetServer connected to 1 - 12 HP

Vectra PCs. PCs, rather than WBTs, were used in the test environment

because the PCs allowed multiple TSE client sessions to be simulated using a

single client system.

To create the network, two sets of six clients were each connected to cascaded

HP J3128A AdvanceStack 10Base-T hubs.

NetServerLHx

10 Mbits

Microsoft Terminal Server

HP NetServer

Microsoft Terminal Server Clients

HP Vectra

HP J31218A HP J3128A

Windows NT

(6 workstations) Windows 95

(6 workstations)

Figure 5

Other manuals for D7171A - NetServer - LPr

117

This manual suits for next models

Table of contents

Other HP Server manuals

HP mt41 User instructions

HP ProLiant Plus User manual

HP D7171A - NetServer - LPr User manual

HP ProLiant ML110 User instructions

HP EX475 - MediaSmart Server - 512 MB RAM User manual

HP EX485 - MediaSmart Server - 2 GB RAM How to use

HP Cluster Platform Hardware Kits v2010 User manual

HP Blade bc1000 Manual

HP Compaq t5730 User manual

HP t5740 - Thin Client User manual

HP ProLiant DL385 Generation 2 (G2) User instructions

HP T5630w - Compaq Thin Client User manual

HP ProLiant DL320e Installation instructions

HP RP7405/7410 User manual

HP LC2000r - NetServer - 128 MB RAM User manual

HP BL35p - ProLiant - 2 GB RAM User manual

HP Compaq t5710 User manual

HP LC2000r - NetServer - 128 MB RAM User manual

HP ProLiant DL380 Quick reference guide

HP 9000 Series User manual

HP ProLiant DL100 G2 DPSS User manual

HP Integrity rx8620 Operating instructions

HP Performance Optimized Datacenter 40c North... Manual

HP NetServer LH3 User manual

Popular Server manuals by other brands

Dell

Dell PowerEdge R410 Getting started

PowerNAS

PowerNAS CMA quick start guide

Supero

Supero FatTwin F617R3-FT user manual

MSI

MSI P1-109N-L70 user manual

Dell

Dell PowerEdge XE9640 Installation and service manual

Dell

Dell PS4100 Hardware

EtroVISION

EtroVISION EV3151 user manual

NEC

NEC Express5800/T110j System configuration guide

Compaq

Compaq ProLiant 3000 Maintenance and service guide

Dell

Dell PowerEdge 3YPMN Getting started guide

iRobo

iRobo IPC2U user manual

Nortel

Nortel 1000 Con?guration guide user manual

Asus

Asus AP7500 Hardware reference guide

Avid Technology

Avid Technology AirSpeed 5000 manual

Milestone

Milestone Husky IVO 350T Getting started

Planet

Planet NAS-7410 user manual

Lenovo

Lenovo ThinkSystem SR635 V3 user guide

Dell

Dell PowerEdge HS5610 Installation and service manual