Full Disclosure Report Exchange MAPI Messaging Benchmark 3 (MMB3) Category: Single Server Hardware: Dell PowerEdge 1955 Software: Exchange Server 2003 (6.5.7226.0)122/ Test Profile: MAPI Messaging Benchmark 3 Date Accepted: 7/28/06 Revision History 7/28/06 original submission Page 1 of 13
Executive Summary Dell PowerEdge 1955 Single Server Test results MMB3 score 12500 Response time 304 milliseconds (ms) CPU utilization 75% Avg. queue 86 Messages submitted Messages delivered Messages sent 529396 (4-hour steady state period) 1311892 (4-hour steady state period) 529359 (4-hour steady state period) Server configuration CPU 3.0 GHz Intel Xeon 5160 Dual Core CPU count 2 RAM L1 cache L2 cache L3 cache Operating system Storage Controller 8GB Instruction: 12KB Data: 8KB 4MB N/A Microsoft Windows 2003 Server Enterprise Edition 1) 2 x 36GB disks for operating system, Active Directory, page file, and Exchange Server system files 2) 360 x 36GB disks for Exchange Information Store and transaction log files 1) Dell SAS 5 controller (operating system, Active Directory, Exchange Server system files) 2) Dell/EMC CX700 and CX500 SAN (Exchange Information Store and transaction log files) NIC One Integrated Intel PRO/1000 MT Gigabit adapter Results based on 4 hours of steady state running. 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. For information on how MMB3 results differ from deployment and configuration information refer to the on Benchmark vs. Production Configuration Disclosure Note section. Test conducted by Dell Performance Analysis Lab using Microsoft Exchange Load Simulator Tool in June 2006. Page 2 of 13
Dell PowerEdge 1955 Dell s Innovative 9th Generation PowerEdge Servers Through ground-breaking hardware design and continued focus on fewer system updates, Dell s 9th generation PowerEdge servers help reduce the complexity involved in managing data, whether you are a large enterprise or a small business. The servers are designed to a Dell-developed Behavioral Specification that defines consistent hardware layout and user interaction across all server models in this and future PowerEdge generations. Featuring the latest Intel Xeon processors, the 9th generation PowerEdge servers offer the power and performance you expect from Dell. The Dell PowerEdge 1955 Server Saves Valuable Data Center Resources By sharing one chassis infrastructure across multiple servers, the Dell PowerEdge 1955 blade server offers exceptional rack density. As many as 10 blade servers fit in one 7U chassis and share components like power, cooling, server management and embedded I/O Pass-Through devices. The result can be reduced data center floor space usage, a 70 percent reduction in necessary cables and simpler installation and maintenance. High Performance and Availability to Maximize Uptime The Dell PowerEdge 1955 server delivers incredible processing power in an efficient, small form factor. Each blade features dual Intel Xeon dual-core processors that provide next generation performance. It also includes Fully Buffered DIMMs for advanced memory and performance benefits, low latency response and high throughput. Embedded I/O connectivity options include redundant Gigabit Ethernet Layer 2 switches and Pass-Throughs, and Fibre Channel switches and Pass-Throughs, as well as Infiniband Pass-Throughs. High availability is another key benefit of the Dell PowerEdge 1955 server. All of its critical support modules including power, cooling and management features are hot-pluggable and redundant. Plus, each blade supports Single Device Data Correction (SDDC) memory technology that enables the system to withstand multi-bit errors. Simplified Management to Reduce Complexity Simplified management is critical in highly-populated data centers. The PowerEdge 1955 server is designed to help reduce deployment, management and maintenance complexity. Each PowerEdge system is shipped with Dell OpenManage IT Administrator and Server Assistant software that guides you through set-up and installation. The PowerEdge blade chassis also includes Dell Remote Access Card (DRAC)/MC management modules and embedded KVMs that enable management from a remote location. Furthermore, the Dell PowerEdge 1955 blade server fits into the current Dell modular server chassis to protect your technology investment and help reduce Total Cost of Ownership. Page 3 of 13
http://www.microsoft.com/windows/catalog/server/default.aspx?subid= 22&xslt=detail&pgn=f8611b24-0dc8-33ef-0b5b-daff4ad1dbce Index EXECUTIVE SUMMARY... 2 INDEX... 4 1 BENCHMARK VS. PRODUCTION CONFIGURATION DISCLOSURE NOTE... 5 2 TEST RESULTS... 6 2.1 RESPONSE TIMES (LATENCIES)... 10 2.2 MESSAGE THROUGHPUT... 10 3 TEST CONFIGURATION... 11 3.1 LOAD GENERATOR CONFIGURATION... 12 4 ADDITIONAL CONFIGURATION AND TUNING... 13 Page 4 of 13
1 Benchmark vs. Production Configuration Disclosure Note This test measures the messaging throughput of a single server, single site topology. Its purpose is to measure the maximum throughput of a Microsoft Exchange Server on this hardware configuration. This can provide a benchmark for comparing hardware and/or software products, but cannot be used as a deployment guide for production environments. For deployment specific information contact a Microsoft or Dell representative. The MMB3 benchmark does not account for: Usage profiles not matching that of the Load Simulator MMB3 profile; Per user storage, and per server backup requirements; Fault tolerance requirements; Anti-virus processes and effects on the server; UBE/UCE (spam) mail flow; Workloads other than MAPI private folder access. This includes Public Folder, NNTP, POP3 and other email interfaces; Multiple Exchange Server deployments, where additional resources are required to forward mail intra-site; Connectors, Links and Replication to remote Exchange sites; Network topologies, bandwidth availability, latency requirement and SLA related factors like QOS and fail-over path issues; Page 5 of 13
2 Test Results The new MAPI Messaging Benchmark (MMB3) measures throughput in terms of a specific profile of user actions, executed over an 8 hour working day. This benchmark is different from the MMB2 setting that was used with Exchange 2000 in that the rate of client requests is significantly greater for the MMB3 profile. Summary Supported Benchmark Load Benchmark Profile Protocol Length of Steady State Length of Test 12500 MMB3s MAPI Messaging Benchmark 3 (MMB3) Exchange MAPI 4 Hours 8 Hours Transactions in Total Total Messages Submitted 529396 Total Message Recipients Delivered 1311892 Total Messages Sent 529359 Message Recipients Delivered / Messages Submitted 2.48 Transaction Load (per hour) Messages Submitted / hour 132358 Message Recipients Delivered / hour 327996 Messages Sent / hour 132349 Transaction Load (per Second) RPC Read Bytes / sec 2340540764 RPC Write Bytes / sec 2127079613 Processor Average Max Min % Processor Time 78 86 71 Database Average Max Min Database cache size 1275068416 1275068416 1000000000 Table opens/sec 2062 2493 1682 Page 6 of 13
Memory Utilization Average Max Min Available MBytes 5496 5611 5423 Cache Faults/sec 1507 2735 731 Free System Page Table Entries 12904 13043 12791 Pages / sec 2 5 1 Pool Nonpaged Bytes (Bytes) 31051871 31383552 30584832 Pool Paged Bytes (Bytes) 49579959 50270208 48582656 System Cache Resident Bytes 140177004 151842816 129888256 Transition Faults/sec 5 61 2 MSExchangeIS Mailbox Average Max Min Folder Opens / sec 48 55 42 Message Opens / sec 143 182 110 MSExchangeIS Send Queue Average Length 98 219 34 MSExchangeIS Receive Queue Average Length 0 0 0 MSExchangeIS Average Max Min Active User Count 154 207 109 RPC Average Latency (ms) 14 33 6 RPC Num. of Slow Packets 1 5 0 RPC Packets/sec 1708 2035 1482 RPC Read bytes/sec 409873 574949 287840 RPC Requests 24 46 7 RPC Operations/sec 2926 3506 2490 RPC Write bytes/sec 7589891 9788458 5957237 TempTable Current MSExchangeIS VM Largest Block Size 494181372 497680384 491192320 MSExchangeIS VM Total 16MB Free Blocks 2 2 2 MSExchangeIS VM Total Free Blocks 275 279 261 MSExchangeIS VM Large Free Block Bytes 522468348 525967360 519479296 Paging File Average Max Min % Usage (_Total) 2 2 2 Page 7 of 13
Processor Utilization Average Max Min System Processor Utilization (%) 75 84 67 System Processor Interrupts/sec (Total) 16669 18428 14881 Process % CPU Time - Store 252 282 224 Process % CPU Time - Inetinfo 8 9 6 Exchange server is also domain controller? (yes/no) Process % CPU Time LSASS (on domain controller) Yes 7 8 5 Handle Count (STORE) 19942 20288 19583 Private Bytes (STORE) 1988962935 2045333504 1,000,000,000 Virtual Bytes (STORE) 2506041836 2512211968 1,000,000,000 Working Set (STORE) 1778851833 1836163072 1,000,000,000 Handle Count (Inetinfo) 4205 4279 4085 Private Bytes (Inetinfo) 40783358 42020864 38408192 Virtual Bytes (Inetinfo) 481461178 484020224 477835264 Working Set (Inetinfo) 2116445665 2174296064 1000000000 SMTP Server Average Max Min Cat: Address lookups completions/sec 131 164 109 Cat: LDAP searches/sec 10 12 9 SMTP Categorizer Queue 0 2 0 DNS Queries/sec 0 0 0 SMTP Local Queue 108 239 38 Messages Currently Undeliverable 0 0 0 Messages Delivered/sec 40 49 32 Messages Received/sec 0 0 0 Messages Sent/sec 0 0 0 NDRs Generated 0 0 0 Remote Queue Length 0 0 0 System Average Max Min System Processor Queue Length 5 23 0 System Context Switches/Sec 36685 41385 32350 Page 8 of 13
Disk Utilization (Aggregate for Database Logical Disks) Average Max Min Logical Drive Utilization (%) 66 66 66 Disk Reads/Sec 11452 13424 9596 Disk Read Bytes/Sec 214152400 261929088 173610832 Disk Writes/Sec 3428 4152 2892 Disk Write Bytes/Sec 87318016 125025024 65370768 Disk Avg. Disk sec / Read 0 0 0 Disk Avg. Disk sec / Write 0 0 0 Average Disk Queue Length 76 104 60 Disk Utilization (Aggregate for Transaction Log Logical Disks) Average Max Min Logical Drive Utilization (%) 1 1 1 Disk Reads/Sec 0 0 0 Disk Read Bytes/Sec 0 0 0 Disk Writes/Sec 774 868 650 Disk Write Bytes/Sec 19535760 32539584 11858600 Disk Avg. Disk sec / Read 0 0 0 Disk Avg. Disk sec / Write 0 0 0 Average Disk Queue Length 0 0 0 Network Utilization Average Max Min Packets Sent/sec 2341 2783 1974 Packets Received/sec 3047 3672 2519 Bytes Sent/sec 3277055 4360640 2426207 Bytes Received/sec 811590 1029612 662615 Page 9 of 13
2.1 Response Times (Latencies) Client Actions 95 th Percentile Response Time (in milliseconds) Send 1015 Read 141 Reply 78 Reply All 79 Forward 94 Move 203 Delete 125 Permanently Delete 141 S+ Free/Busy 156 Browse Calendar 219 Make Appointment 1234 Request Meeting 1656 Create Smart Folder 219 Delete Smart Folder 500 Create Rule 172 Delete Rule 219 Apply View/Sort 6078 Weighted Total 304 2.2 Message Throughput Summary of the MMB3 profile for an 8 hour day: Expected Measured Messages Submitted/MMB3/Day 85.0 84.7 Messages Delivered/MMB3/Day 210 209.9 Average Recipients per Message 2.47 2.48 Page 10 of 13
3 Test Configuration Describe below the configuration of the Exchange Server machines (physical) used for this test. If more then one, they should have an identical configuration. Hardware Exchange Server Domain Controller (if remote) Vendor Dell n/a Model PowerEdge 1955 n/a Processor Intel Dual Core Xeon 3.0GHz n/a # of Processors (Physical) 2 n/a # of Processors (Logical) 4 n/a Hyper-Threading enabled? N/A n/a Primary Cache Instruction: 12KB Data: 8KB n/a Secondary Cache 4MB n/a Other Cache N/A n/a Memory 8GB n/a Disk Subsystem Disk Controllers Other Hardware 1) 2 x 36GB disks for operating system, Active Directory and Exchange Server system files 2) 360 x 36GB disks for Exchange Information Store and transaction log files 1) Dell SAS 5 controller (operating system, Active Directory and Exchange Server system files 2) Dell/EMC CX700 and CX500 SAN (Exchange Server Information Store and transaction log files) Integrated Intel PRO/1000 MT Gigabit adapter n/a n/a n/a Page 11 of 13
Mail Software Exchange Server Domain Controller (if remote) Vendor Microsoft Corporation n/a Mail Server Exchange Server n/a Release Version 2003 Service Pack 1 n/a Operating System Exchange Server Domain Controller (if remote) OS Version Microsoft Windows 2003 Enterprise Edition n/a Service Pack Service Pack 1 n/a OS Hot-fixes/patches n/a n/a File System Type NTFS n/a Network Exchange Server Domain Controller (if remote) Type of Network Ethernet n/a Network Speed 1 Gbit n/a TCP/IP Offload/Checksum Tx/Rx TCP/IP Checksum Tx/Rx TCP Segmentation n/a PCI Flow Control? Generate and Respond n/a Interrupt Coalescing? Hardware Default n/a 3.1 Load Generator Configuration This section holds all the configuration parameters of the load generator machines used in the test. # of Load Generators (LG) 18 Total # of LG processes 12500 Simulated Users/Process 16 clients @ 736 users 1 client @ 625 users 1 client @ 99 users Model PowerEdge 1650 Processor 1.4 GHz Intel Pentium III # of Processors (Physical) 1 (/numproc=1) # of Processors (Logical) 1 Hyper-Threading enabled? N/A Page 12 of 13
Memory Network Controller Network Bandwidth Operating System 1GB Embedded Intel PRO/1000 XT 1 Gbit Windows XP Professional SP1 KB331320 Microsoft Outlook 2003 4 Additional Configuration and Tuning Describe below in items any modifications done to the Exchange Server(s) and the server/client operating systems. These modifications include but are not restricted to performance tuning changes like registry keys and boot.ini settings. All modifications must be approved by Microsoft prior to the testing and submission of the MMB3 result. Boot.ini /3GB /userva=3030 Registry HeapDeCommitFreeBlockThreshold=0x00040000 Show Advanced Counters=0x1 Active Directory msexcheseparamcachesizemax=307200 Dell, the Dell logo, and PowerEdge are trademarks of Dell Inc. Microsoft and Windows are registered trademarks of Microsoft Corporation. Other trademarks and trade names may be used in this document to refer to either the entities claiming the marks and names or their products. Dell disclaims proprietary interest in the marks and names of others. Copyright 2006 Dell Inc. All rights reserved. Reproduction in any manner whatsoever without the express written permission of Dell Inc. is strictly forbidden. For more information contact Dell. Page 13 of 13