Compaq 850R - ProLiant - 32 MB RAM Manuel utilisateur



November 1998

ECG082/0898

Prepared by Messaging and

Collaboration Solutions Business

Unit

Compaq Computer Corporation

Contents

Compaq ProLiant 7000............... 3

Exchange Server

Performance Test Results.......... 3

What the Benchmarks Don’t

Tell You ..................................... 4

MAPI Messaging Benchmark

(MMB) - LoadSim Medium

User Redefined ......................... 4

MMB Transaction Load............. 4

Message Mix Description.......... 5

Related Documents .................... 9





 !!!

""#

Abstract: Compaq achieved record-breaking Exchange Server

scalability by achieving 15,000 MMB on a ProLiant 7000 server

equipped with four 400-MHz Intel Pentium



II Xeon



processors.

Using Microsoft’s Load Simulation utility, the ProLiant 7000 with

four 400-MHz processors was tested at Compaq’s Microsoft

Competency Center in Redmond, Washington. The favorable results

of this testing demonstrate the superior performance capabilities of

four-processor servers based on the Intel Pentium



II Xeon



processor as compared to eight-processor servers based on the Intel

Pentium



Pro available from competitors.

Compaq and Microsoft have worked to optimize Microsoft Exchange

Server performance on Compaq server products to provide an

optimal balance between performance, availability, manageability,

and cost. Compaq not only provides world-class server platforms,

but also the experience necessary for the successful deployment of

messaging and collaborative applications

 ! "

ECG082/0898

Notice

The information in this publication is subject to change without notice and is provided “AS IS” WITHOUT

WARRANTY OF ANY KIND. THE ENTIRE RISK ARISING OUT OF THE USE OF THIS

INFORMATION REMAINS WITH RECIPIENT. IN NO EVENT SHALL COMPAQ BE LIABLE FOR

ANY DIRECT, CONSEQUENTIAL, INCIDENTAL, SPECIAL, PUNITIVE OR OTHER DAMAGES

WHATSOEVER (INCLUDING WITHOUT LIMITATION, DAMAGES FOR LOSS OF BUSINESS

PROFITS, BUSINESS INTERRUPTION OR LOSS OF BUSINESS INFORMATION), EVEN IF

COMPAQ HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.

The limited warranties for Compaq products are exclusively set forth in the documentation accompanying

such products. Nothing herein should be construed as constituting a further or additional warranty.

This publication does not constitute an endorsement of the product or products that were tested. The

configuration or configurations tested or described may or may not be the only available solution. This test

is not a determination or product quality or correctness, nor does it ensure compliance with any federal

state or local requirements.

Product names mentioned herein may be trademarks and/or registered trademarks of their respective

companies.

Compaq, Contura, Deskpro, Fastart, Compaq Insight Manager, LTE, PageMarq, Systempro, Systempro/LT,

ProLiant, TwinTray, ROMPaq, LicensePaq, QVision, SLT, ProLinea, SmartStart, NetFlex, DirectPlus,

QuickFind, RemotePaq, BackPaq, TechPaq, SpeedPaq, QuickBack, PaqFax, Presario, SilentCool,

CompaqCare (design), Aero, SmartStation, MiniStation, and PaqRap, registered United States Patent and

Trademark Office.

Netelligent, Armada, Cruiser, Concerto, QuickChoice, ProSignia, Systempro/XL, Net1, LTE Elite,

Vocalyst, PageMate, SoftPaq, FirstPaq, SolutionPaq, EasyPoint, EZ Help, MaxLight, MultiLock,

QuickBlank, QuickLock, UltraView, Innovate logo, Wonder Tools logo in black/white and color, and

Compaq PC Card Solution logo are trademarks and/or service marks of Compaq Computer Corporation.

Microsoft, Windows, Windows NT, Windows NT Server and Workstation, Microsoft SQL Server for

Windows NT are trademarks and/or registered trademarks of Microsoft Corporation.

Intel, Pentium, Pentium II, Xeon, and Pentium Pro are trademarks and/or registered trademarks of Intel

Corporation.

Cheyenne, ArcServe and Computer Associates are the trademarks and/or registered trademarks of

Computer Associates, Inc.

NetWare and Novell are registered trademarks and intraNetWare, NDS, and Novell Directory Services are

trademarks of Novell, Inc.

Enterprise-Class Microsoft Exchange Server Scalability on Compaq ProLiant 7000 Pentium II Xeon

Servers

Solutions Guide prepared by Messaging and Collaboration Solutions Business Unit

First Edition (November 1998)

Document Number ECG082/0898

 ! #

ECG082/0898

Compaq ProLiant 7000

The Compaq ProLiant 7000 is the ultimate standards-based server delivering the most scalable

performance and the highest levels of availability. The ProLiant 7000 supports up to four

PentiumII Xeonprocessors and 8 GB of system memory. Its architecture supports Enhanced

PCI Hot-Plug and a Triple-Peer PCI bus design (five 64-bit PCI, four 32-bit PCI, and one ISA

slot). Standard with the ProLiant 7000 is the 64-bit PCI SMART 3100ES Array Controller that

supports 3 Wide Ultra SCSI-3 channels and 64 MB of cache. Leveraging Fibre Channel, the

Compaq ProLiant 7000 can support nearly 7 TB of external storage. Combined with the latest

high-availability features and processor technology, the ProLiant 7000 is designed for the most

demanding and mission-critical applications.

Exchange Server Performance Test Results

The tests were conducted using Microsoft MAPI Messaging Benchmark (MMB). The MAPI

Messaging Benchmark (MMB) measures throughput in terms of a specific profile of user actions,

executed over an 8 hour working day. This benchmark utilizes the ‘Medium User’ setting of the

Load Simulator MAPI tool. Results should be interpreted as a benchmark for messaging

throughput and should not be confused with deployment recommendations. Factors such as

backup/restore, topology and other issues should be considered when planning a deployment.

Table 1. Performance Highlights (ProLiant 7000, (4) PentiumII Xeon400-MHz, 4 GB RAM)

MMB 15,000

Response Time (milliseconds) 343

Messages Submitted (8-hour period) 213,021

Messages Delivered (8-hour period) 260,707

Message Recipients Delivered (8-hour period) 1,179,107

Messages Sent (8-hour period) 57,315

Note: Complete Test results disclosure can be found later in this document.

Table 2. Configuration of the tested Compaq ProLiant 7000

Compaq ProLiant 7000 Tested Configuration

(4) PentiumII Xeon400-MHz – 1 MB L2 cache per processor

4 GB RAM

(1) SMART 3100ES Array Controller with 64 MB Cache (configured for 100% WB)

OS/Pagefile/Exchange DS/MTA Files/Exchange Log Files: (2) 4.3-GB Drives – RAID1

Exchange Information Store Files: (16) 4.3-GB Drives – RAID0

Compaq Netelligent (100BaseTX) network interface card (NIC)

Windows NT Server v4.0 Enterprise Edition (using /3GB BOOT.INI switch)

Exchange Server v5.5 – Enterprise Edition

Note: Complete Test results disclosure can be found later in this document.

 ! $

ECG082/0898

What the Benchmarks Don’t Tell You

It is important to understand that benchmarks such as these are designed to give Exchange Server

implementation planners baseline references for understanding the capabilities of hardware

platforms from a single vendor such as Compaq or other competing hardware vendors. When

interpreting these benchmarks, two things should be kept in mind.

First, consider whether benchmarks are performed on what can be referred to as customer-

deployable configurations. A hardware vendor may publish a result that is based on a platform or

configuration that should not be deployed in a real-world Exchange Server deployment. For

example, many vendors (including Compaq) publish results using disk subsystems configured

with RAID0. While RAID0 does provide the highest levels of disk subsystem performance, it

fails to provide any protection against data loss. Compaq recommends deploying an Exchange

Server with disk fault tolerance such as RAID1 or RAID5 for the highest levels of data

protection.

Second, most vendors, including Compaq, conduct benchmarks for Exchange Server that are

single-server in nature. Also, keep in mind that benchmarks do not account for issues such as

backup and disaster recovery or information store maintenance sizing. Whatever the issue, care

must be taken when interpreting benchmarks to ensure that they represent useful information for

your Exchange Server deployment and are based on valid simulation methodologies.

While it is significant that the ProLiant 7000 server can successfully scale to 15,000 medium

MAPI e-mail users (MMB) in a single-server benchmark exercise, Compaq recommends careful

evaluation of all issues involved in real-world Exchange Server deployments – issues such as

management, administration, and disaster recovery.

MAPI Messaging Benchmark (MMB) - LoadSim Medium

User Redefined

To distinguish clearly between throughput benchmarks and capacity planning information for

Microsoft Exchange Server, Microsoft has established the MAPI Messaging Benchmark (MMB)

based on the workload from LoadSim Medium User profile. The MAPI Messaging Benchmark

representative workload focuses on the resulting throughput and clearly communicates the profile

under test.

The workload profile has not changed from the former LoadSim Medium User, but expressed in

clearer fashion. The intent is to make sure that customers can understand the MAPI Messaging

Benchmark workload and can compare it to other platforms. In addition, the renaming of the

benchmark reinforces that the test is a measurement of messaging throughput and that additional

considerations are required in capacity planning.

MMB Transaction Load

The transaction load created by the benchmark is equivalent to the user actions over an eight-hour

day as shown in Table 3.

 ! %

ECG082/0898

Table 3. Average User Actions

User Action Actions Per Day

Check Inbox 12

Send Message 14.18

Avg. Recipients per Message 4.7

Messages Received 66.3

Read Message 81.3

Move Message 16.3

Delete Message 32.5

Update Calendar 5

30% of all mail messages have one distribution list recipient. Average size of DL 10 recipients.

(Recipients created by distribution lists are included in the summary above.) All users are logged

on prior to the benchmark as they are assumed to be using mail in a corporate setting. Mail is not

cleared from the deleted items folder during the test as this is assumed to happen when the user

logs off.

Message Mix Description

The weights used when the Load Simulator randomly selects which message to send are listed in

the Table 4.

Table 4. Message Mix Description

Message Files Body Attachment Content Description Weight

Ups1k.msg  Body as RTF 60

Ups2k.msg  Body as RTF 16

Ups4k.msg  Body as RTF 4

Ups10kat.msg  10K Body as RTF - Notepad attachment 6

Upsxlatt.msg  15K Body as RTF - Microsoft Excel spreadsheet attached 4

Upswdatt.msg  16K Body as RTF - Microsoft Word document attached 4

Upsbmobj.msg  43K Body as RTF - Bitmap attachment 2

Load Simulator

The main tool used in generating the workload the MMB benchmark was Microsoft Load

Simulator. Load Simulator is a tool for simulating a client user load on an Exchange Server. Its

purpose is to enable a single Windows NT machine called a LoadSim client to simulate multiple

Microsoft Exchange client users.

The operation of Load Simulator users is governed by a Load Simulator profile. This profile

controls factors such as how long a Load Simulator "day" is, how many e-mail messages to send

in a day’s time, how many times to open and read e-mail, whether to use distribution lists,

whether to use public folders, etc.

 ! &

ECG082/0898

Load Simulator creates a highly accurate simulation of reality. It mimics the full Microsoft

Exchange Client in many respects. First, it uses .MSG files, the same format used by the

Exchange Client. This guarantees that messages generated by Load Simulator have the same

properties as those sent by real users of the Exchange Client. Second, Load Simulator uses the

same MAPI remote procedure call (RPC) semantics as those used by the Client. Third, Load

Simulator registers MAPI change notifications in the same manner as they are registered by the

Client. Finally, Load Simulator even emulates the Microsoft Exchange Client list box cache,

which the Client uses for folder and message panes in the viewer when a user browses and selects

messages on the server. For more information on LoadSim Medium canonical profiles, refer to

the LoadSim documentation at http://www.microsoft.com/exchange/library/loadsim55x86.exe.

Table 5. TEST DISCLOSURE

LoadSim Clients Configuration

Network Topology (100Base T, Token Ring, etc.) 100 Base-TX

Number and type of clients (≤40) 2x5/133, 128 MB RAM (≤400 users each)

or better (indicates minimum configuration)

Number and type of hubs/concentrators (full

duplex, switching, etc.) Compaq Netelligent 5708 Switch and Netelligent

2624 Hubs

Number of clients/segment 20

Client CPU type and speed in percentages 2P/133-MHz Pentiumprocessors or better

Client network controller broken down by

percentages Compaq Netelligent 10/100

Client network software name and version (drivers,

protocols, redirector) Microsoft Windows NT Workstation 4.0 with SP3

TCP/IP

Size of any client network cache None

Network controller software Compaq Netelligent 10/100 driver

LoadSim version 5.5 (Build 2187)

Note: Response time measurements were taken from a LoadSim Control Client simulating 100

users configured with 96 MB RAM and a Pentium/166 CPU. The client is located on an isolated

network segment connected to a 100-Mb/s switch.

Benchmark Tuning

Table 6 shows tuning parameters for Exchange Server that varied from default settings for

purposes of the benchmark exercise. Compaq recommends using the default settings specified by

the Microsoft Exchange Performance Optimizer for production server deployments.

Table 6. Benchmark Tuning

Parameter Default Setting Benchmark Setting

IS Buffers (4K each) 524,000 375,000

Max Open Tables 32,000 200,000

IS Send Threads 4 8

 ! 

ECG082/0898

IS Deliver Threads 4 8

MTA Submit/Deliver Threads 2 8

MTA TCP Threads 2 6

MTA TCP Control Blocks 20 25

Table 7. Version Information (Server)

Name Type Version or Date

Compaq NT SSD Software/Drivers 2.09

SMART ECI Driver Software/Drivers 1.06

System Firmware ROM 7/8/98

SMART 3100ES Firmware ROM 2.72

Windows NT Server EE Operating System 4.0 SP3

Exchange Server EE Application 5.5 (Build 1960.8)

Table 8. Performance Data

15,000 MMB (Measured during test run at steady state)

Summary

Supported Benchmark Load 15,000 MMB

Weighted 95th-Percentile Score 343 milliseconds

Benchmark Profile MMB (Medium MAPI User)

Protocol Exchange MAPI

Length of Steady State 4 hours

Length of Test 8 hours

Transaction Load (hourly)

Messages Submitted 26,627.69

Message Recipients Delivered 147,388.46

Messages Sent 7,164.49

Transaction Load (per Second)

Message Opens/Sec 53.75

Folder Opens/Sec 14.05

RPC Read Bytes/Sec 47,250.31

RPC Write Bytes/Sec 383,111.15

Transaction Queues

IS Send Queue Average Length 5.50

MTA Work Queue Average Length 8.04

 ! '

ECG082/0898

Processor Utilization

System Processor Utilization (%) 83.74%

System Processor Queue Length 4.29

System Context Switches/Sec 5438

Process % CPU Time – Store 272.33%

Process % CPU Time – DS 25.38%

Process % CPU Time – MTA 15.91%

Memory Utilization

Available Bytes 1838 MB

Pages/Sec 0.015

Process Working Set Bytes – Store 1.99 GB

Process Virtual Bytes – Store 2.78 GB

Logical Drive Utilization

IS Database Disk Reads/Sec 409.40

IS Database Disk Writes/Sec 244.01

IS Database Average Disk Queue

Length 11.36

IS Log Disk Reads/Sec 0.93

IS Log Disk Writes/Sec 159.82

IS Log Average Disk Queue Length 0.34

Note: Performance results were measured using Microsoft NT Performance Monitor.

Measurements were obtained by measuring averages for the period of steady-state activity (i.e.

after 15, 000 users were successfully logged on). Tests measure the messaging throughput of a

single-server, single-site topology. For deployment-specific information contact a Microsoft or

Compaq representative. More information can be found at:

http://www.microsoft.com/exchange/support/deployment/planning/deploy.asp?A=5&B=1

Table 9. User Response Times (Latencies) from Load Simulator

Client Actions 95th-Percentile Response Time (in Milliseconds)

Send 531

Delete 421

Move 550

Submit 441

 ! (

ECG082/0898

Manuels Serveur populaires d'autres marques

HP AlphaServer ES47 Manuel utilisateur

Dell

Dell PowerEdge R410 Manuel utilisateur

PowerNAS

PowerNAS CMA Manuel utilisateur

Supero

Supero FatTwin F617R3-FT Manuel utilisateur

MSI

MSI P1-109N-L70 Manuel utilisateur

Dell

Dell PowerEdge XE9640 Manuel

Dell

Dell PS4100 Instructions de montage

EtroVISION

EtroVISION EV3151 Manuel utilisateur

NEC

NEC Express5800/T110j Manuel utilisateur

HP ProLiant Gen9 Manuel utilisateur

Dell

Dell PowerEdge 3YPMN Manuel utilisateur

iRobo

iRobo IPC2U Manuel utilisateur

Nortel

Nortel 1000 Con?guration guide Manuel utilisateur

Asus

Asus AP7500 Manuel d'utilisation et d'entretien

Avid Technology

Avid Technology AirSpeed 5000 Manuel utilisateur

HP Integrity rx2600 Guide d'installation

Milestone

Milestone Husky IVO 350T Manuel utilisateur

Planet

Planet NAS-7410 Manuel utilisateur