-
The EEC has been working closely with Hitachi Data Systems to grow the HDS footprint and provide more resources for our customers. This had led to an agreement to add two of Hitachi’s BladeSymphony2000 systems, the first of which is scheduled to arrive in February. The BladeSymphony adds a level of power and flexibility that the EEC needs in order to support the ever increasing demands of our customers test scenarios by bringing Hitachi’s mainframe technology to the x64 server world. Through the use of their Virtage technology, we’ll have the ability to create logical partitions within the blade chassis allowing the EEC to scale from 12 to 96 cores in a compact 10U enclosure.
For more on BladeSymphony: http://bladesymphony.com/products/bladesymphony_2000/index.html
Steve Cole
Sr. Business Development Manager - EEC
-
As you know the Windows HPC team utilizes the EEC for product benchmarking, ISV engagement and customer validation. During SC09, Vince Medillo, Sr. Director HPC here at Microsoft, released some benchmarking data that was gathered on the Rihannon cluster that showed the Windows HPC Server 2008 beta rivaled that of Linux.

To read more about the announcement, please read Edward J. Correia's article at Network Computing.
http://www.networkcomputing.com/servers-storage/microsoft-windows-hpc-beta-on-par-with-linux.php
Kevin Engman
Product Manager - EEC
-
SC09 is going well and the work by the ASU students is very impressive. To learn more about it, please check our Wenming Ye's post on the Windows HPC Survival Guide blog. http://blogs.msdn.com/hpctrekker/archive/2009/11/14/super-computing-2009-next-week.aspx
Kevin Engman
Product Manager - EEC
-
This week we are attending SC09 http://sc09.supercomputing.org/ educating folks about the EEC and supporting the HPC team. They will be demoing Excel Runner, which is tested at the EEC. We are also excited for the Arizona State University's work here in the student challange. They are the only team that is running on Windows HPC Server 2008 and spent 2 weeks in October at the EEC validating their appications.
If you're at SC09, please stop by booth 1135 and say hello.
Kevin Engman
Program Manager - EEC
-
A great new book on the new features and refinements in Windows Server 2008 R2 is available in the Microsoft download center. It’s called Introducing Windows Server 2008 R2 and is targeted primarily at Windows server administrators who are responsible for hands-on deployment and day-to-day management of Windows-based servers for large organizations.
The EEC hosted dozens of customers for product validation testing during the development cycle for R2 and we are excited to see it finally available for companies to deploy into their environments.
The direct link to the eBook is:
http://download.microsoft.com/download/5/C/0/5C0BD0AB-040D-4C56-A60B-661001012DDA/Windows_Server_2008_R2_e-book.pdf
Kevin Engman
Product Manager - EEC
-
During a recent engagement with Siemens SWAT team we were able to benchmark Fusion IO cards in their exchange environment and used JetStress 2010. The server used was an HP 385 G5 running Windows 2008 R2.
Below are the storage performance results from testing:
|
I/O Database Reads Average Latency (msec) |
4.830 |
|
I/O Database Writes Average Latency (msec) |
12.094 |
|
I/O Database Reads per second |
7,761,885 |
|
I/O Database Writes per second |
6,406.448 |
|
I/O Database Reads Average Bytes |
32,869.093 |
|
I/O Database Writes Average Bytes |
35,203.116 |
|
Total Database Read Throughput |
243.31 MB/sec |
|
Total Database Write Throughput |
215.08 MB/sec |
We look forward to increased usage and availability of next generation storage removing the historical bottleneck of storage IO.
-Dustin Clarkson
Program Manager II - EEC
-
Last week 5 Samba developers once again came to the EEC for to do some interop work. This time around they were going to be working with the Directory Replication Service (DRS) with the aim of moving from 'Samba4 to Samba4' replication to 'Samba4 and AD
replication'. The results were a big success. To learn more about it, please read Andrew Bartlett's account. http://people.samba.org/people/2009/10/05
Kevin Engman
Product Manager - EEC
-
This week, we’ve been in the EEC doing our scale testing on the world’s largest PKI, issuing 100s of millions of certificates from 100s of CAs to devices around the world. The entire design is built on WS08R2 Hyper-V and WS08/WS08R2 CAs.
To simulate one of their Hyper-V hosts, we used a similar machine to the ones being used currently in the hosting facilities: a Sun Microsystems 2.4GHz, 4 socket, quad core machine with 64GB. We loaded our host with 10 VMs, each assigned a single VCPU and 6GB. All 10 of these VMs were connected to an nCipher netHSM 2000. To generate load, each CA VM was paired with a single DC and 5 client machines, each assigned a single VCPU and 2GB and separated from the CA by a WAN simulator that added latency and throughput constraints based on the VPNs used to link bases to the hosting facility. We used the EntGenReq tool to have each client machine open 4 request sessions and requests 1000000 2K key certs per session.
After <24 hours, we had issued >20 million certificates from this single physical chassis. During these tests, we found that:
-
Per VM CPU load was ~25%, total host CPU load was ~20%
-
Relatively little memory was required by the CA VMs, even at this high stress; thus we’re optimizing the design to increase the density of CA VMs per chassis, to 30:1 (2GB per VM)
-
The performance bottleneck in this design is the HSM; as we increased the number of CA VMs being stressed, our requests per second per CA fell significantly, from >100 to ~18-20, giving a net issuance rate for the entire chassis of ~200 per second
-
When investigating the HSM, it became clear that it was the gating component (150 request queue remained saturated and CPU is pegged at consistently at 85%)
Overall, this testing was a great validation of the performance of ADCS. Microsoft software ran as fast as the HSM would allow and we gracefully handled response delays introduced by it. Also, the fact that we’re able to run this configuration entirely on Hyper-V and get ~30 CAs per physical host provides an efficient scale story for even the very largest and most complex environments around.
Dustin Clarkson,
Program Manager - EEC
-
During one of our engagements with Tuev Nord, (one of the world’s largest technical service providers of the Mobility, Industrial Services, International, Natural Resources and Training and Human Resources fields), the Windows Server Hyper-V feature team supplied an exhaustive list of websites covering documentation and tools around Hyper-V.
I thought I’d share that with you.
TechNet On Hyper-V
Ø Hyper-V Getting Started Guide
Ø Hyper-V: Using Hyper-V and Failover Clustering
Ø Hyper-V: Using Live Migration with Cluster Shared Volumes in Windows Server 2008 R2
Ø Planning for Disks and Storage
Ø Configuring Disks and Storage
Performance Counter Information
Ø Performance Tuning Guidelines For Server 2008 R2 (whole section on Hyper-V)
Ø Monitoring Hyper-V Performance
Ø Hyper-V Performance Counters – Part one of many - The overview
Ø Hyper-V Performance Counters – Part two of many – “Hyper-V Hypervisor” counter set
Ø Hyper-V Performance Counters – Part three of many – “Hyper-V Hypervisor Logical Processors” counter set
Ø Hyper-V Performance Counters – Part four of many – “Hyper-V Hypervisor Virtual Processor” and “Hyper-V Hypervisor Root Virtual Processor” counter set
Performance and Load Tools
Ø IOMeter
o IOMeter (http://www.iometer.org/) Pretty much the industry standard for storage benchmarks. See Tony’s link for some good profiles (http://blogs.msdn.com/tvoellm/archive/2009/05/07/useful-io-profiles-for-simulating-various-workloads.aspx)
Ø NTTCP
o http://www.microsoft.com/whdc/device/network/TCP_tool.mspx
Ø WCAT
o http://www.iis.net/downloads/default.aspx?tabid=34&g=6&i=1466 – Microsoft Web Capacity Analysis Tool
Ø BizTalk
o Measuring Performance on Hyper-V (http://msdn.microsoft.com/en-us/library/cc768535(BTS.10).aspx
o Tools for Measuring Performance (http://msdn.microsoft.com/en-us/library/cc768530(BTS.10).aspx
Ø Exchange Servers
o Exchange Server Stress and Performance Tool http://www.microsoft.com/downloads/details.aspx?familyid=B9E200D5-18B9-4734-B9D9-9E6EFC48BCB6&displaylang=en
Ø Active Directory Performance Tool
o AdTest http://www.microsoft.com/downloads/details.aspx?familyid=4814FE3F-92CE-4871-B8A4-99F98B3F4338&displaylang=en
Ø Geekbench
o http://www.primatelabs.ca/geekbench - Pretty new but seems well respected.
Ø IOZone
o http://www.iozone.org/
Ø SQL
o Running SQL Server 2008 in a Hyper-V Environment - Best Practices and Performance Recommendations.
Ø SPEC Benchmarks
o http://www.spec.org/benchmarks.html - Pretty widely used they have CPU, Graphics, Mail, Java, Etc… I am not a huge fan and they’re a bit pricy.
Ø TPC
o http://www.tpc.org – Standard for Databases you can get TPC-E for free but it’s kind of a pain to setup.
Kevin Engman
Product Manager - EEC
-
For those folks that like facility news, yesterday Microsoft’s Managed Lab Services team gave a small group of reporters and bloggers the first external tour of the their new data center in Redmond, aka Redmond Ridge 1.
Artist’s rendering of Redmond Ridge 1

Todd Bishop, editor & founder of TechFlash, was one of the lucky few and he talked about his experience and the facility on his blog. http://www.techflash.com/seattle/2009/09/microsoft_shifting_server_labs_from_offices_to_remote_green_facility.html
I’m looking forward to taking a tour myself, especially since we just upgraded our facility. It will be interesting to see how their infrastructure is laid out. It’s always fun to see a new building, especially after having just finished a construction on one of your own.
Kevin Engman
Product Manager - EEC
-
Last week 20 new Dell™ PowerEdge™ R710 servers arrived and are quickly being brought online for a customer engagement occurring in the next couple weeks. These servers contain Quad-Core Intel® Xeon® X5550 CPU’s, 24GB memory, eight 300GB SAS HDDs, and much more. We are pleased to have this new hardware available for customers and ISV’s as they validate our products and services.
The new Dell servers in POD 2 lit up by the IRIS system

To learn more about Dell’s PowerEdge R710 server line, please visit:
http://www.dell.com/us/en/enterprise/servers/server-poweredge-r710/pd.aspx?refid=server-poweredge-r710&s=biz&cs=555
Kevin Engman,
Product Manager - EEC
-
Recently I was able to work with BWIN, the world’s largest online sports-betting and gaming provider, at the EEC while they tested the scaling of their workload on the world’s largest Windows Server, a HP Superdome with 128 cores and 256 threads, and HP’s largest disk array, a HP XP24000.
We were able to test various configurations (64 cores, 128 cores, Hyperthreading on or off) during their stay and gathered tons of performance data around optimizing BWIN’s SQL Server performance. This was the first real-life customer test on a system with more than 64 processors, a great proof point for Windows Server 2008 R2 and SQL Server 2008 R2.
Check out the screen shot, task manager showing 256 processors!
To find out more about the Superdome go to
http://h20341.www2.hp.com/integrity/cache/342431-0-0-0-121.html
Details about the XP24000 can be found at http://h18006.www1.hp.com/products/storageworks/xp24000/specs.html
Gunter Zink,
Performance Architect, HP
-
While I was on vacation a couple weeks ago we received 10 new HP ProLiant® DL380 G6's equipped with dual Intel® Xeon® X5570 CPU’s, , 24GB memory, dual port 8Gbps fiber channel adapter, 2 146GB 15K SAS drives and dual power supplies. These were the 1st G6’s that the EEC has obtained and they were quickly added to a customer’s test environment while they continued their validation of SQL Server 2008 R2 and Windows Server 2008 R2.
This week two more DL380 G6’s will be arriving. These will have dual Intel® Xeon® E5530 CPU’s, 24GB memory, 2 146GB 15k SAS drives and dual power supplies, bringing the total number of DL380 G6’s available for customer and product group testing to 12 .
We’d like to thank HP again for their support of the EEC and our customers. Without the support of our partners, we would not be able to provide the value we do to the validation process.
For more information on HP’s DL380 G6 server series please visit the following HP sites.
DL380 G6 product page:
http://h10010.www1.hp.com/wwpc/us/en/en/WF25a/15351-15351-3328412-241644-241475-3884082.html
DL380 G6 Datasheet:
http://h18004.www1.hp.com/products/quickspecs/DS_00147/DS_00147.pdf
DL380 G6 Quickspecs:
http://h18004.www1.hp.com/products/quickspecs/13234_na/13234_na.html
Kevin Engman
Product Manager - EEC
-
This week we just completed installing a new IBM XIV® Storage System: Model 2810 in san-alley. Ours is configured with 180 1TB SATA drives, and 15 modules; a data module consists of one Intel® Xeon® Quad-core 64-bit CPU, 8 GB fully buffered DIMM RAM. The flexibility that this hardware provides will greatly increase the effectiveness of the EEC by allowing us to meet the demands of our customer and product group engagements.
Hardware Specifications:
|
Hardware Highlights |
|
Total Storage Central Processing Units |
15 Intel® Xeon® 64-bit quad-core |
|
Combined Physical Memory |
120 GB |
|
Raw Storage |
180 |
|
Disk to Cache Bandwidth |
240 Gbps |
|
Internal Switching Capacity |
160 Gbps |
|
Fibre Channel Ports (Speed: 1 Gbs, 2 Gbs, & 4 Gbs) |
24 |
|
iSCSI Gigabit Ethernet Ports |
6 |
|
UPS |
3 |
|
Gigabit Ethernet Switch |
2 |
|
Disk Hot Spares |
3 |
|
Data Module Hot Spare |
1 |
|
Max Snapshots |
16,000 |
Some of the main new features of the XIV are that they provide increased performance, more reliability/protection, thin provisioning, lower power consumption and the ability to take snapshots. For more information on IBM XIV® Storage System visit the IBM website at: http://www.xivstorage.com.
If you’d like to visit the EEC and validate our products using the XIV, please visit http://www.microsoft.com/eec.
Kevin Engman
Product Manager - EEC
-
Earlier this year the Office Communications Group (OCG) approached the EEC and our Partners (HP, F5 and Shunra) to help them build an environment where they could perform the necessary testing to create this much needed white paper. This was a challenging project, but was completed because of the wonderful collaboration amongst the teams from each company.
Here’s an overview of the white paper:
This white paper describes a site resiliency solution for Office Communications Server 2007 R2. The solution includes an Enterprise pool that spans two geographically separate sites. The solution provides a failover mechanism between the two sites to ensure that Office Communications Server functionality such as instant messaging, presence, and conferencing remains available even if one of the sites becomes unavailable.
This white paper is divided into three main sections:
· The Solution section provides an overview of the tested and supported site resiliency solution described in this paper.
· The Test Methodology section describes the testing topology, expected behavior, and test results.
· The Findings and Recommendations section provides practical guidance for deploying your own failover solution.
You can download the white paper here:
http://www.microsoft.com/downloads/details.aspx?displaylang=en&FamilyID=c930febb-3a44-4bf3-969d-1c52675a7063
Thanks to all those that participated.
Kevin Engman
Product Manager - EEC