IBM Lotus Domino 8.5 server performance, part 1: IBM Lotus Notes performance

Increasing value to the customer is an ongoing theme for the IBM® Lotus® Domino® server. The growing volume of information in today’s enterprise strains existing hardware infrastructures. Lotus Domino 8.5 contains features that help relieve the stress on the expensive processor and storage subsystems. This article shows the reductions in processor and disk utilization with Lotus Domino 8.5 and Lotus Notes® clients.

Yang Bin, Software Performance Analyst, IBM

Yang Bin is currently a member of the IBM Lotus Domino for IBM i team with a focus on Lotus Domino performance. You can reach Yang Bin at yangbin@cn.ibm.com.



Eric Bjorklund, Software Performance Analyst, IBM

Eric Bjorklund is a member of the Power Family performance team, focusing on IBM Lotus Domino for IBM i. You can reach him at erbjor@us.ibm.com.



Rich Buck, Software Engineer, IBM

Rich Buck is a member of the Lotus Domino performance team, with primary focus on Lotus Domino for Sun Solaris and Lotus Domino for Microsoft Windows performance. You can reach him at richbuck@us.ibm.com. He tested and wrote the Solaris section of this article.



Wu W Huang, Software Engineer, IBM

Wu W Huang is a member of the Lotus Domino Performance team, with primary focus iNotes and XPages on Windows and IBM Lotus Domino performance on System Z . You can reach Wu Huang at wuhuang@us.ibm.com. He tested performance of the Classic and XPages discussion database.



Angelo Lynn, Software Engineer, IBM

Angelo Lynn is an engineer on the Lotus Domino performance team. His current focus is Lotus Domino performance on Windows-based operating systems. He is a recent graduate from Northeastern University. You can reach him at anglynn@us.ibm.com.



Andy Nolet, Software Engineer, IBM

Andy Nolet has been working with customers on Lotus Notes performance-related issues since the late 1990s. Before joining the Lotus Domino performance team, Andy worked for Lotus Support. You can reach him at anolet@us.ibm.com.



Jim Powers, Software Engineer, IBM

Jim Powers is a member of the Lotus Domino performance team. Previously, Jim led the performance team for the Lotus Domino Support organization. His experience with computer systems goes back over 30 years; performing various hardware and software roles throughout his career. You can reach him at jhp@us.ibm.com.



Nirmala Venkatraman, Performance Architect, IBM

Nirmala Venkatraman is a Performance Architect on the Lotus Domino server performance team. You can reach her at nvenkatr@us.ibm.com.



28 July 2009 (First published 03 March 2009)

Also available in Chinese Russian Japanese Portuguese

Editor's note: Know a lot about this topic? Want to share your expertise? Participate in the IBM Lotus software wiki program today.

Introduction

This article is part 1 of a three-part series on performance topics for Lotus Domino 8.5. Be sure to read these developerWorks® articles:

This article presents data that shows that upgrading to Lotus Domino 8.5 should be a positive experience. Yoou can realize substantial I/O reductions, reductions in disk operations per second, and reductions in disk bytes transferred per second. In addition, processor utilization has been reduced by as much as 20 percent. Upgrading to Lotus Domino 8.5 can lower your Lotus Domino deployment total cost of ownership.

Figure 1 shows the general trend of performance improvements achieved with Lotus Domino 8.5 across a selection of platforms with servers of varying capacities.

Figure 1. Server resource reduction with Lotus Domino 8.5 with 4000 simulated Lotus Notes users
Server resource reduction with Lotus Domino 8.5 with 4000 simulated Lotus Notes users

The measurements in this article use Notesbench performance workloads that mimic the behavior of thousands of Lotus Notes clients performing common messaging and calendar operations using a single Lotus Domino server. Two workloads were used; N8Mail simulates the API calls generated by the Lotus Notes 8.0 client, and N85Mail performs the same high-level operations, but with the API calls generated by a Lotus Notes 8.5 client.

In general, the tests were set up with 5000 users defined in the Lotus Domino directory on the server. At the beginning of the test, each user has a mail file that is roughly 256 MB of uncompressed documents, 3000 messages in the inbox, and roughly 380 messages in the trash. The messages in the trash are set to expire and to be deleted at the rate of two messages every 15 minutes throughout the measurement. These tests have transaction logging enabled with the favor runtime setting, and mail journaling is set to journal all messages locally. Domino domain monitoring (DDM) probes are enabled for messaging and operating systems, and all users have mail rules that block mail from ten users external to the test.

With Lotus Domino 8.5, the test environment enabled document compression on the mail databases, which reduced their size from approximately 250 MB to approximately 170 MB. In addition, the test environment enabled Domino attachment and object service (DAOS) after the mail databases were created, and we also enabled DAOS on the mail boxes and journal files.

For the most part, the tests used the default settings of the Lotus Domino server and mail template. A few settings, though, were changed to enhance performance and enable document compression and DAOS on version 8.5. The changes used on all operating systems are listed in table 1, with any additional platform-specific items in the section that discussines results on the platform. See the appendix at the end of this article for more details on what the notes.ini settings shown in table 1 can accomplish.

Table 1. Lotus Domino server configuration for testing on all operating systems
ConfigurationLotus Domino 8.0 measurementsLotus Domino 8.5 measurements
Mail templateMail8.ntfMail85.ntf
WorkloadN8MailN85Mail
Transaction logEnabled / Favor runtimeEnabled / Favor runtime
DAOSNot applicableEnabled with encryption
Mail journalingAll messages localAll messages local
Additional, nondefault, mail database optionsDon’t overwrite free space
Disable automatic update of views
Don’t overwrite free space
Disable automatic update of views
Compress document data
Use Lotus Domino attachment and object service
Server tasks Replica,Router,Update,AMgr,Adminp,Sched,CalConn,RnRMgr,LDAPReplica,Router,Update,AMgr,Adminp,Sched,CalConn,RnRMgr,LDAP
Notes.ini additions NLCache_Size=67108864
Server_Pool_Tasks=80
Server_Max_Concurrent_Trans=100
Server_Show_Performance=1
RouterDbCacheSize=6100
Schedule_No_Validate=1
NSF_DBcache_Maxentries=5100
Create_R8_Databases=1
Debug_NSF_Show_Allstats=1
Same as Lotus Domino 8.0 plus:
Create_R85_Databases=1
Debug_NSF_Compress_All_Notes=1
Create_R85_Log=1
Enable_LZ1_Encrypted_Notes=1

NOTE: The results in this article come from benchmarks executed in a controlled environment. While some effort was made during the creation of the benchmark to include typical user operations, it is likely that real users make different use of Lotus Domino than the narrow range of function that is tested by the benchmark. These numbers should therefore be used primarily to understand the relative performance of the Lotus Domino releases, and they do not represent recommendations for real-world deployment. For assistance with capacity planning, consult your hardware vendor and IBM TechLine.

Also, although we show results on a variety of hardware platforms, these configurations are not of uniform capacity. It is our intent here to focus on the performance of Lotus Domino itself, and this data should not be used to compare platforms against each other.

The following sections show the detailed results on the platforms that we tested.


AIX V6.1 results

Table 2 summarizes the configuration used for the IBM AIX® V6.1 testing.

Table 2. AIX V6.1 configuration
ModelIBM POWER® 570 (9117-MMA)
Processors for test / speed 1 physical and 1 logical 3.5 GHz processor
Memory16 GB
Active physical drivesIBM FastT Fibre Channel storage, six trays each with 14 15000 RPM using RAID 0;
six trays are configured into one logical volume
Active logical volumesOne logical volume for Lotus Domino binaries/data, one local drive for transaction logging
Operating systemAIX V6.1 maintenance level 2 64-bit
Lotus Domino versions Lotus Domino 8.0: 32-bit application
Lotus Domino 8.5: 64-bit application
Notes.ini settings in addition to those commonly used for these testsD85 only NSF_Buffer_Pool_Size_MB=512
Server_Transinfo_range=42
Server_Pool_Tasks=100
NSF_DBCACHE_CLEAN_HOLD_TIME=9999

Test hardware was an IBM Power 570 (9117-MMA) system, which is based on POWER6™ processor-based technology, configured into two LPARs (logical partitions). The test LPAR was configured with a single physical processor and a single logical processor. This configuration was capped and did not use entitlement for additional processor resources. Capping the available unit's processor power allowed measurement of the load at a higher level than would be possible with the variability in processor allocated to the LPAR using virtualized processors and entitlement.

Storage was configured using an IBM DS4000 Series (FastT) into a single logical file system that supported the Lotus Domino binaries, Lotus Domino data files, mail journal database, and the DAOS NLO files. We used the AIX Advanced Journal File system (JFS2) configured with RAID 0. RAID 0 was used as the only files left intact between test runs of the Lotus Domino binaries and installed data files. Mail database files, log files, Mail*.Box, and DAOS files are recreated prior to starting a test. This recreation is done to ensure that every test is started with the databases in the same state. The transaction log files are recreated for every test and located on dedicated storage configured using RAID 0.

The Lotus Domino 8.0 tests used a default NSF buffer pool size of 512 MB; for the Lotus Domino 8.5 tests the NSF buffer pool size was explicitly set to be the same (512 MB). NOTE: The default size for Lotus Domino 8.0 and Lotus Domino 8.5 32-bit Lotus Domino is 512 MB; for 64-bit Lotus Domino 8.5 the default NSF buffer pool size is 1 GB when there is more than 4 GB RAM in the system. Network access was through a single 1 GB Ethernet adapter running in full duplex mode.

Figures 2 and 3 are the resource utilization charts for the two tests at various user loads.

Figure 2. AIX: Percent processor busy
AIX – Percent processor busy

By running the same workload on the same hardware, Lotus Domino 8.5 uses less processor resources at all measured virtual user levels. The processor improvement ranged between 11 percent and 20 percent across the four comparison points. Less is better.

Figure 3. AIX: Total disk operation/second
AIX – Total disk operation/second

Lotus Domino 8.5 does fewer total disk operations across the measured virtual users. The total disk I/O operations improvement ranged between 23 percent and 30 percent across the four comparison points. Less is better.

Figure 4. AIX: Total disk megabytes transferred/second
AIX – Total disk megabytes transferred/second

Figure 4 shows that Lotus Domino 8.5 sends less data as measured by total disk megabytes transferred/second across the measured virtual users. The total disk megabytes transferred improvement ranged between 45 percent and 50 percent across the four comparison points. Table 4 summarizes the resource usage for AIX V6.1. Less is better.

Table 3. AIX V6.1: Resource usage at 4000 users
ResourceLotus Domino 8 Lotus Domino 8.5% change
Processor percent busy3528-20%
Total disk operations/second941701-25%
Total disk reads/second303118-61%
Total disk writes/second638581-9%
Total disk megabytes transferred/second15847%
Total disk megabytes read/second31-67%
Total disk megabytes written/second 127-42%
Shared memory used in MB1,7991,8845%
Process memory used in MB110101 class="numeric">-8%
Network megabytes / second2,312,3172,173,282-6%

Lotus Domino 8.5 showed a reduction in all measured resource requirements, with the exception of shared memory. Note that Lotus Domino 8.5 was the 64-bit application version, and internal data structures are sized and managed differently than with 32-bit application code.

For more details, refer to the developerWorks® Lotus article, "IBM Lotus Domino 8.0.1 for 64-bit server performance."


IBM i results

All performance test results documented here were completed using an IBM i System® 570 with two active processor cores and 8 GB of memory. The system was configured with 48 disk drives with RAID-5 protection. Network access occurred through a single 100 MB Ethernet adapter running in full-duplex mode.

Table 4. IBM i configuration
ModelIBM i570 (9406-MMA)
Processors for test / speed Two 4.7 GHz processor cores
Memory3 GB dedicated for machine pool
5 GB dedicated for base pool
Disk drives48 drives with RAID-5 protection
Active logical volumesOne logical volume for Lotus Domino binaries/data, one local drive for transaction logging
Operating systemIBM i 5.4
Lotus Domino versions Lotus Domino 8.0
Lotus Domino 8.5
Notes.ini settings in addition to those commonly used for these testsNone

The system was configured with one Lotus Domino partition. Lotus Domino 8.0 was tested with N8Mail workload using the Lotus Domino 8.0 mail template. Lotus Domino 8.5 was tested with N85Mail workload using the Lotus Domino 8.5 mail template. The transaction logging and mail journaling files were under the Lotus Domino data directory. Each test was run with 1000, 2000, 3000, and 4000 simulated users, and performance information was collected at each data point.

Figure 5. IBM i: Percent processor busy
IBM i: Percent processor busy

Figure 5 shows the percent processor busy rate at different simulated user loads. As you can see from this data, Lotus Domino 8.5 reduced the processor utilization of the system at each data point. The processor improvement ranged between 11 percent and 15 percent across the four comparison points.

Figure 6. IBM i: Total disk operations/second
IBM i: Total disk operations/second

Figure 6 shows the total disk I/O operations per second at different simulated user loads. As you can see from this data, Lotus Domino 8.5 also reduced the disk I/O operations of the system at each data point. The total disk I/O operations improvement ranged between 22 percent and 32 percent across the four comparison points.

Figure 7. IBM i: Total disk megabytes transferred/second
IBM i: Total disk megabytes transferred/second

Figure 7 shows the total disk MBs transferred per second at different simulated user loads. As you can see from this data, Lotus Domino 8.5 reduced the disk MBs transferred at each data point. The improvement in total disk MBs transferred ranged between 32 percent and 43 percent across the four comparison points.

Table 5. IBM i: Resource usage at 4000 users
ResourceLotus Domino 8 Lotus Domino 8.5% change
Processor percent busy2724-11%
Total disk operations/second14051096-22%
Total disk reads/second838599-29%
Total disk writes/second567496-13%
Total disk MBs transferred/second20.914.3-32%
Total disk MBs read/second12.47.8-37%
Total disk MBs written/second 8.46.5-23%
Base pool page faults/second774572-26%
Network MBs / second2.342.17-7%

Our lab data, summarized in table 5, showed that Lotus Domino 8.5 reduced the processor utilization of the system at 4000 users. The page faulting rate and number of disk I/O operations and disk MBs transferred were also reduced including disk reads and writes. We also saw a reduction in network utilization with Lotus Domino 8.5. Overall, Lotus Domino 8.5 showed a better performance on processor utilization, disk I/O operations, disk MBs transferred, and network utilization.


Linux results

This section covers the Notes Remote Procedure Call results of testing on an Intel® system running 64-bit Linux® version SuSE SLES 10 x86/64. It should also be noted that even though the operating system is 64-bit, the Lotus Domino versions used in the testing were all 32-bit. There are advantages to running a 32-bit application such as Lotus Domino on 64-bit Linux. One of these advantages is memory. On 32-bit Linux, a 32-bit application can get a maximum of 3 GB of memory to use, whereas on 64-bit Linux a 32-bit application can get up to 4 GB of memory. Lotus Domino is good at utilizing this additional memory for additional server tasks or applications.

The test server used for this set of tests is an Intel Xeon® MP dual-core processor, which has 8 GB of RAM and three DS4000 disk subsystems (IBM FastT) configured as nine Raid 0 logical units. For the tests, the simulated mail user databases are spread evenly over eight of the logical units. A single drive logical unit on one of the DS 4000s is used for the Lotus Domino transaction log files. Table 6 summarizes the details of the Linux configuration.

Table 6. Linux configuration
ModelIntel 64-bit platform
Processors for test / speedIntel Xeon MP configured as two cores /3400 MHz
Memory8 GB
Active physical drivesIBM FastT Fiber Channel storage, three trays using RAID 0
3 trays are configured into nine logical volumes
Active logical volumesEight logical volumes for Lotus Domino data, one local drive for transaction logging
Operating systemSLES 10 x86-64
Lotus Domino versions Lotus Domino 8.0: 32-bit application
Lotus Domino 8.5: 32-bit application
Notes.ini settings in addition to those commonly used for these testsMEM_AddressableMemSizeMB=3500
ConstrainedSHMSizeMB=3000
MEM_EnablePreAlloc=1
NSF_buffer_pool_size_MB=512

The notes.ini parameters that follow were used in addition to those mentioned previously. These notes.ini parameters should not be considered default Linux settings, and they should be used only with complete understanding of what they do. For most systems, the default Lotus Domino settings are the best choice.

MEM_AddressableMemSizeMB=3500 ConstrainedSHMSizeMB=3000 MEM_EnablePreAlloc=1

These server notes.ini parameters deal with the allocation of memory for the Lotus Domino server. As mentioned previously, a system running 64-bit Linux can allocate 4 GB of memory to a 32-bit application. The notes.ini parameter MEM_AddressableMemSizeMB=3500 tells Lotus Domino that it has 3.5 GB of memory to use. This statement sounds like a contradiction because we just said that a 32-bit application can get 4 GB of memory. The operating system needs some of this 4 GB for mapping and other memory pools for all the Lotus Domino tasks that are running. Usually 500 MB is ample for this task.

The second notes.ini parameter, ConstrainedSHMSizeMB=3000, tells Lotus Domino how much shared memory it can use. Because most of the memory that Lotus Domino requires is shared memory, we used 3 GB for these tests. This value might not be suitable in your environment. You need to know how much shared memory and how much process memory are required in your environment to run all your applications first.

The third notes.ini parameter, MEM_EnablePreAlloc=1, tells Lotus Domino to preallocate the shared memory as defined by the second notes.ini parameter. This approach helps prevent Lotus Domino from crashing with an insufficient shared memory error, but it is a balancing act because it’s not wise to lock down more shared memory than needed. Doing this lockdown of shared memory could cause a shortage of local memory for applications to use and cause both insufficient memory errors and application failures.

The following figures show a comparison of Lotus Domino 8.0 and Lotus Domino 8.5 running the workloads described in the introduction section.

Figure 8. Linux: Percent processor busy
Linux: Percent processor busy

Figure 8 shows the processor savings across the board that were attained with Lotus Domino 8.5. The processor improvement ranged between 15 percent and 19 percent across the four comparison points.

Figure 9. Linux: Total disk operations/second
Linux: Total disk operations/second

Figure 9 shows the I/O savings that were attained with Lotus Domino 8.5. The total disk I/O operations improvement ranged between 29 percent and 37 percent across the four comparison points.

Figure 10. Linux: Total disk megabytes transferred/second
Linux: Total disk megabytes transferred/second

Figure 10 displayed the reduction in I/O data that was attained with Lotus Domino 8.5. The total disk megabytes transferred improvement ranged between 42 percent and 50 percent across the four comparison points. Table 7 summarizes the resource usage at 4000 users.

Table 7. Linux - Resource usage at 4000 users
ResourceLotus Domino 8 Lotus Domino 8.5% change
Processor percent busy4839-19%
Total disk operations/second1363969-29%
Total disk reads/second760391-49%
Total disk writes/second603578-4%
Total disk MBs transferred/second179-47%
Total disk MBs read/second72-71%
Total disk MBs written/second 107-30%
Network MBs / second2,406,9182,252,393-6%

Lotus Domino 8.5 shows a reduction in all measurements when compared to Lotus Domino 8.0. Memory consumption is not shown in table 7 because it was preallocated for both runs by notes.ini parameters.


Solaris results

The server used for the Sun Solaris testing is detailed in table 8. It is a four- processor domain with six RAID 0 logical units of nine drives each to hold the Lotus Domino data and executable files. In addition, two file systems on a seventh logical unit contain the transaction log and the DAOS objects.

Table 8. Solaris configuration
ModelSun 6800
Processors for test / speedFour / 1050 MHz
Memory32 GB
Active physical drives54
Active logical volumesSeven – T3 logical volumes configured as RAID 0
Operating systemSolaris 10
Lotus Domino versions Lotus Domino 8.0 – 32-bit application
Lotus Domino 8.5 – 32-bit application
Notes.ini settings in addition to those commonly used for these testsNone

This system was configured with the Lotus Domino settings outlined at the beginning of this article, and the NSF buffer pool was allowed to default to 512 MB for both measurements. This system had an abundance of memory that was put to good use by Solaris for file caching. Also, the file systems were using the normal Sun 8 K block size, which resulted in somewhat different disk I/O rates than those that we saw for the other systems that use a 4 K block size. We saw excellent improvements in processor and disk usage with Lotus Domino 8.5 on this system. See figures 11 and 12.

Figure 11. Solaris – Percent processor busy
Solaris – Percent processor busy

Comparing the processor utilization of Lotus Domino 8.5 with that of Lotus Domino 8.0, there is up to a 20 percent (relative) reduction on Solaris.

Figure 12. Solaris: Total disk operations/second
Solaris: Total disk operations/second

Reducing the disk I/O operations translates to real savings as fewer spindles are required to support a certain level of user activity. Here, with Lotus Domino 8.5 on Solaris we are seeing a 30 percent reduction.

Figure 13. Solaris: Total disk megabytes transferred/second
Solaris: Total disk megabytes transferred/second

In addition to the reduction in disk I/O operations, we see a dramatic reduction in the MBs transferred, over 60 percent in most cases, as shown in figure 13.

Table 9 summarizes the resource usage for Solaris.

Table 9. Solaris 10: Resource usage at 4000 users
ResourceLotus Domino 8 Lotus Domino 8.5% change
Processor percent busy7057-19%
Total disk operations/second862605-30%
Total disk reads/second14247-67%
Total disk writes/second719558-22%
Total disk MBs transferred/second28.369.29-67%
Total disk MBs read/second15.551.28-92%
Total disk MBs written/second 12.818.01-37%
Shared memory used in MB1,2391,2834%
Process memory used in MB218225 class="numeric">3%
Network MBs / second2.111.97-7%

On this Solaris server, we saw tremendous I/O and processor benefits when moving to Lotus Domino 8.5. Disk operations are reduced 30 percent, and there is a processor savings of 19 percent when compared to Lotus Domino 8.0. With Lotus Domino 8.5, you can store both the design and message content in a compressed format on disk. This approach has the added benefit of enhancing the operating system file cache, in effect doubling its efficiency because it now contains mostly compressed data. This system, which has an exceptionally large 32 GB of memory, shows that effect particularly well. The number of read operations that needed to go to the physical disks is reduced by 67 percent, while the bytes needed to be read from disk is down 92 percent compared to Lotus Domino 8.0.


Microsoft Windows 2003 server results

All performance test results documented here were completed using an IBM xSystem® 3850 with four active processor cores running Microsoft® Windows® 2003 Standard x64 Edition. The system has 8 GB of installed memory. The system was configured with 42 disk drives with RAID 0. Network access occurred through a single 100 MB Ethernet adapter running in full-duplex mode. Table 10 summarizes the details of the Windows 2003 configuration.

Table 10. Windows 2003 configuration
ModelIBM x3850-[8863MC1]
Processors for test / speedFour processorsat 3.6 GHz
Memory8 GB installed
Active physical drives42 disks
Active logical volumesSeven logical volumes of RAID 0
Operating systemWindows 2003 Standard x64 Edition
Lotus Domino versions Lotus Domino 8.0: 32-bit application
Lotus Domino 8.5: 32-bit application
Notes.ini settings in addition to those commonly used for these testsNone

The system was configured with one Lotus Domino 32-bit partition. The Lotus Domino 64-bit version for the Windows 64-bit operating system is also available, but for this article the performance team used the 32-bit version of Lotus Domino. Lotus Domino 8 was tested with the N8Mail workload using the Lotus Domino 8 mail template (mail8.ntf). Lotus Domino 8.5 was tested with N85Mail workload using the Lotus Domino 8.5 mail template (mail85.ntf). DS4000 disk subsystems (IBM FAStT 600) configured as seven RAID 0 logical units were used. The Lotus Domino executable files were installed on one 200 GB logical unit. The mail databases were spread across three logical units, each 800 GB in size. Transaction logging was set up on a separate three-disk logical unit. Each test was run with 1000, 2000, 3000, and 4000 simulated users, and performance information was collected at each data point for comparison of Lotus Domino 8 and Lotus Domino 8.5.

Figure 14. Windows 64: Percent processor busy
Windows 64: Percent processor busy

Figure 14 shows the percent processor busy rate at different simulated user loads. As you can see from this data, Lotus Domino 8.5 reduced the processor utilization of the system at each data point. The processor improvement ranged between 17 percent and 39 percent across the four comparison points.

Figure 15. Windows 64: Total processor operations/second
Windows 64: Total processor operations/second

Figure 15 shows the total disk I/O operations per second at different simulated user loads. As you can see from this data, Lotus Domino 8.5 also reduced the disk I/O operations of the system at each data point. The total disk I/O operations improvement ranged between 24 percent and 33 percent across the four comparison points.

Figure 16. Windows 64: Total disk megabytes transferred/second
Windows 64: Total disk megabytes transferred/second

Figure 16 shows the total disk MBs transferred per second at different simulated user loads. As you can see from this data, Lotus Domino 8.5 reduced the disk MBs transferred of the system at each data point. The total disk MBs transferred improvement ranged between 33 percent and 44 percent across the four comparison points.

Table 11 summarizes the resource usage for Windows 2003.

Table 11. Windows 2003 server: Resource usage at 4000 users
ResourceLotus Domino 8 Lotus Domino 8.5% change
Processor percent busy3421-38%
Total disk operations/second1020687-33%
Total disk reads/second402157-61%
Total disk writes/second619530-14%
Total disk MBs transferred/second12.47.0-43%
Total disk MBs read/second3.30.9-73%
Total disk MBs written/second 9.16.1-33%
Shared memory used in MB1,3211,313-0.6%
Process memory used in MB50.151.8 class="numeric">3%
Network MBs / second2.322.16-7%

Our testing produced a processor percent and an I/O percent decrease when running Lotus Domino 8.5 at 4000 users when compared to Lotus Domino 8.0. Overall, Lotus Domino 8.5 produces a better performance on processor utilization, disk I/O operations, disk megabytes transferred, and network utilization. With these performance gains by deploying Lotus Domino 8.5 for Windows, customers can enjoy cost-saving advantages that might lead to server consolidation.


Linux on System z results

The 64-bit Lotus Domino 8.5 on Linux System z® is finally here. For this article, we focus on using the new N8Mail and N85Mail workloads with a maximum of 4000 users to compare the performance results in Lotus Domino 8.5 and Lotus Domino 8.0 on Linux System z. One new feature for Lotus Domino 8.5 is the Lotus Notes document compression. This feature significantly reduces I/O and disk space usage. On System z, though, if Lotus Notes document compression is not enabled there is a larger processor utilization improvement and an I/O improvement. Lotus Domino 8.5 comes with Lotus Notes document compression disabled by default. Both configurations and results are detailed below.

All performance test results documented here come from one LPAR on a zSeries® z9® model 2096-S07 Business Class. This LPAR enabled two processors, configured 12 GB memory, and a single Lotus Domino Partition running an instance of the Lotus Domino server. This system booted SLES 10 with few updates. We used a single GB Ethernet open system architecture card. Our LAN was isolated. All disks were allocated from an IBM System DS8300 array with each disk configured as a 3390 model 9. Client mail databases were distributed evenly over 64 LVM files systems, each allocated across four volumes in a single LVM, providing 28 GB of usable space per file system. Separate volumes were allocated for the Lotus Domino execution, Lotus Notes data, Lotus Domino address book, mailbox 1 to 8, and transaction logging. Another four volumes were used in an LVM for mail journaling, and four volumes were used in an LVM for DAOS (Domino attachment and object service). The EXT3 file system was used on Linux for System z. Table 12 summarizes the Linux hardware configuration.

Table 12. Linux hardware configuration
ModelZ9 2096-S07
Processors for test / speedTwo processors (2096)
Memory12 GB
Enterprise storage serverDS8300
Active logical volumes64 LVM mail databases (four volumes each)
Two LVM mail journaling and mail DAOS (four volumes each)
Seven volumes for Notes data, names, mailbox, transaction log, and Notes bin
Operating systemSLES 10 (64-bit)
Lotus Domino versions Lotus Domino 8.0: 32-bit application
Lotus Domino 8.5: 32-bit application
Notes.ini settings in addition to those commonly used for these testsNSF_Buffer_Pool_Size_MB=386
MailCompactDisabled=1
NSF_DBcache_maxentries=6000
NLCACHE_SIZE=104857600

Lotus Domino 8.5 No Notes Document Compression only
NSF_COMPRESS_TXN_LOGS=1
DEBUG_NSF_COMPRESS_ALL_NOTES=0 (Default)

Figures 17 and 18 show the Lotus Domino 8 with the mail8 template run with N8Mail, Lotus Domino 8.5 with the mail85 template run with N85Mail, and the Lotus Domino 8.5 with no Lotus Notes document compression with mail85 template run with N85Mail. Each interval showed the average of the one hour steady state of each 1000-user increment.

Figure 17. zLinux: Percent processor busy
zLinux: Percent processor busy

Figure 17 shows a processor cost ranging from 0 percent to 6 percent degradation for Lotus Domino 8.5 and Lotus Domino 8 with Lotus Notes document compression enabled. Figure 17 shows processor improvement ranging from 11 percent to 13 percent for Lotus Domino 85 with Lotus Notes compression disabled (default for Lotus Domino 8.5) and Lotus Domino 8.

Figure 18. zLinux: Total disk operations/second
zLinux: Total disk operations/second

The disk operation per second shows I/O improvement ranging from 27 percent to 33 percent for Lotus Domino 8.5 compared with Lotus Domino 8; see figure 18. It also shows improvement for Lotus Domino 8.5 with Lotus Notes document compression disabled, ranging from 16 percent to 24 percent compared with Lotus Domino 8.

Figure 19. zLinux: Total disk megabytes transferred/second
zLinux: Total disk megabytes transferred/second

The total disk MBs transferred per second, detailed in figure 19, shows a huge improvement ranging from 40 percent to 43 percent for Lotus Domino 8.5 compared with Lotus Domino 8. It also shows improvement with Lotus Notes document compression disabled ranging from 29 percent to 33 percent for Lotus Domino 8.5 compared with Lotus Domino 8.

Table 13 summarizes the resource usage for Linux on System z.

Table 13. Linux on System z, with Lotus Notes document compression: Resource usage at 4000 users
With Louts Notes document compressionLotus Domino 8 Lotus Domino 8.5% change
Processor percent busy61645%
Total disk operations/second1030751-27%
Total disk reads/second361166-54%
Total disk writes/second670584-13%
Total disk MBs transferred/second11.46.8-40%
Total disk MBs read/second1.40.65-54%
Total disk MBs written/second 106.2-38%
Shared memory used in MB1,5161,5603%
Process memory used in MB235255 class="numeric">9%
Network MBs / second2.262.11-7%

At 4000 users, table14 shows that the total I/O operation improved 27 percent and the total I/O bytes transferred improved 40 percent for Lotus Domino 8.5 compared with Lotus Domino 8. Total shared memory increased 3 percent, and total process memory increased 9 percent. Total network bytes received and transmitted improved 7 percent.

Table 14. Linux on System z, without Lotus Notes document compression: Resource usage at 4000 users
Without Louts Notes document compressionLotus Domino 8 Lotus Domino 8.5% change
Processor percent busy6153-13%
Total disk operations/second1030854-17%
Total disk reads/second361234-35%
Total disk writes/second670619-8%
Total disk MBs transferred/second11.48.1-29%
Total disk MBs read/second1.410.65-54%
Total disk MBs written/second 107.2-28%
Shared memory used in MB1,5161,5603%
Process memory used in MB235263 class="numeric">12%
Network MBs / second2.262.1-7%

At 4000 users, table 14 shows that the total I/O operation improved 17 percent and the total I/O bytes transferred improved 29 percent for Lotus Domino 8.5 and Lotus Domino 8. Total shared memory increased 3 percent, and total process memory increased 12 percent. Total network bytes received and transmitted increased 7 percent.

In conclusion, Lotus Domino 8.5 is a 64-bit application on zLinux. Lab measurements indicated that minor processor costs up to 6 percent for Lotus Domino 8.5 if Lotus Notes document compression is enabled, but with great I/O improvement, up to 33 percent total disk operations and up to 43 percent of total disk bytes transferred for Lotus Domino 8.5 compared with Lotus Domino 8. Lotus Domino 8.5 comes with Lotus Notes document compression disabled by default. With the default, the processor improved up to 13 percent; total disk operation improved up to 24 percent; total disk bytes transferred improved up to 32 percent. The processor and the I/O reductions translate into a greatly reduced total cost of ownership for Lotus Domino 8.5.


Conclusion

From the measurements reported in this article, upgrading to Lotus Domino 8.5 is a positive experience. You can realize substantial I/O reductions, reductions in disk operations per second by 22 percent to 33 percent, and reductions in disk bytes transferred per second by 31 percent to 67 percent. In addition, processor utilization has been reduced by as much as 20 percent. Upgrading to Lotus Domino 8.5 can lower your Lotus Domino deployment total cost of ownership.

Appendix. Explanation of Notes.ini settings
Notes.ini parameter nameExplanation
Create_R8_Databases Applies to Lotus Domino 8 servers. Set it to 1 to create new ODS48 databases on Lotus Domino 8 servers.
Create_R85_Databases Applies to Lotus Domino 8.5 servers. Set it to 1 to create new ODS51 databases on Lotus Domino 8.5 servers.
Create_R85_Log Applies to Lotus Domino 8.5 servers. Set it to create the Lotus Domino transaction log in Lotus Domino 8.5 format.
Debug_NSF_Compress_All_Notes Set it to 1 to enable Document compression for all databases on the Lotus Domino server.
NSF_COMPRESS_TXN_LOGS Set it to 1 to enable compression of log records (used by Linux on System z).
DEBUG_ENABLE_SYS_V_SHMSet it to 1 to let Lotus Domino use System V for shared memory instead of map files on UNIX® platforms.
DEBUG_NSF_SHOW_ALLSTATS New in Lotus Domino 8. Set it to 1 to enable Lotus Domino I/O statistics collection for database metadata, views, and objects.
EVENT_CORRELATION_POOL_SIZESpecifies the size of the event correlation pool used by the event task and DDM.
Enable_LZ1_Encrypted_NotesSet it to 1 to enable LZ1 compression of encrypted documents in the Mail journal database.
EVENT_POOL_SIZE Set this variable to specify the amount of memory used for unprocessed events, unprocessed notifications, and event suppression. The default setting is 5242880 (5 MB).
log_mailroutingSpecifies the level of logging of router events performed by the router process.
MEM_AddressableMemSizeMB Set to override the default setting for the amount of memory Lotus Domino sees as addressable memory (where VALUE is in megabytes, and the default varies by platform).
MEM_EnablePreAllocSet it to 1 to enable preallocation of shared memory.
NLCACHE_SIZESpecifies the size of the namelookup cache in bytes. The default value is 16 MB.
NSF_Buffer_Pool_Size_MBSpecifies the maximum size (in bytes) of the NSF buffer pool, a section of memory dedicated to buffering I/O transfers between Lotus Domino and disk storage.
NSF_Dbcache_Maxentries Specifies the number of databases that a server can hold in its database cache at one time.
RouterDbCacheSize Specifies the size of the router database cache used to cache mail databases in the router process.
SCHEDULE_NO_VALIDATE Enables or disables whether SchedMgr validates its busytime database entry on a daily basis, as follows:
0 (default): Enables validation
1: Disables validation
Server_MAX_CONCURRENT_TRANSSets the limit for the number of concurrently scheduled transactions on a server.
Server_Pool_Tasks Total number of physical threads in the Lotus Domino server (DbServer) threadpool.
Server_Show_PerformanceApplies to servers. Specifies whether server performance events are displayed on the console. Set this variable to 1 to display server performance events on console.
ServerTasks Specifies the tasks that begin automatically at server startup and continue until the server is shut down.

Resources

Comments

developerWorks: Sign in

Required fields are indicated with an asterisk (*).


Need an IBM ID?
Forgot your IBM ID?


Forgot your password?
Change your password

By clicking Submit, you agree to the developerWorks terms of use.

 


The first time you sign into developerWorks, a profile is created for you. Information in your profile (your name, country/region, and company name) is displayed to the public and will accompany any content you post, unless you opt to hide your company name. You may update your IBM account at any time.

All information submitted is secure.

Choose your display name



The first time you sign in to developerWorks, a profile is created for you, so you need to choose a display name. Your display name accompanies the content you post on developerWorks.

Please choose a display name between 3-31 characters. Your display name must be unique in the developerWorks community and should not be your email address for privacy reasons.

Required fields are indicated with an asterisk (*).

(Must be between 3 – 31 characters.)

By clicking Submit, you agree to the developerWorks terms of use.

 


All information submitted is secure.

Dig deeper into IBM collaboration and social software on developerWorks


static.content.url=http://www.ibm.com/developerworks/js/artrating/
SITE_ID=1
Zone=Lotus
ArticleID=373647
ArticleTitle=IBM Lotus Domino 8.5 server performance, part 1: IBM Lotus Notes performance
publish-date=07282009