While much of the discussion and attention will be focused on the very high end VDI scaling cases next week, I thought it would be useful to show some stuff at the other extreme. We also have been testing in the 125-250-500-1000 user ranges on our NS20, which is an mid-range unified system. This is something someone could get started in the $30K range. Note - I'm not talking about what we're doing at the other end of the extreme - the 10K+ use case (we'll talk about that next week).
There is a detailed document that covers these best practices, H5719: "EMC Infrastructure for Deploying VDI in the Enterprise". The document is on powerlink, and will be up on EMC.com shortly. The full validation test report showing all the performance envelopes will be up right after VMworld on powerlink.
UPDATE: Folks, the NetApp side has made some comments that in their view we're not being clear here: http://blogs.netapp.com/virtualstorageguy/ Vaughn is a good guy, so I'm going to take the personal slag (ouch - used car salesman) in stride. We are replicating the LUN objects in this case, and they are the unit of replication. If you will have more than roughly 200 VDI VMs per cluster, you need to have several VMs per LUN. The PowerVDI tool automates this. These ratios mean it's more accuate to say "you will invariably achieve 100:1 space savings, and may achieve more depending on the scaling goal". (the ratio of source VMs per source LUN needs to be higher at higher scales, and therefore 100:1 is the worst case model). There is NOTHING intrinsically wrong with NetApp's approach, there is NOTHING intrinsically wrong with EMC's approach. Each have advantages and disadvantages, and are a function of each vendor trying to solve the customer problem. The ratio of LUNs to VMs is not 1:1. We don't do this because we love block - while we can take 96 filesystem snapshots, only 16 of those are writeable, so this approach is better for EMC customers. Just setting the technical record straight.
If you want to see some interesting performance metrics - read on!
One thing we've been wondering about is the cache/non cached response. We've been putting it through it's paces in a lot of use cases in the creation of the docs, here's one example. Keep in mind, this is the lower-end of the spectrum - targeted at simple and easy.
BTW - in some of these tests, we are HAMMERING the config. In the chart below, we tested to 500 VMs in a single 4+1 RAID group.
A few things stand out - the initial iterations, before the cache was loaded, was distinctly different than the post-cached response, but man, you would have a grumpy end user when it takes 60 seconds to open excel the first time.
Here's what it looks like at 250 users per 4+1 RAID group.
Much better. It's amazing the dropoff in response time to complete the task as the array caches the workload.
What about memory oversubscription?
This is fascinating, and highlights the incredible power of memory dedupe. The ESX servers we were testing had 32GB of RAM, and 8 cores. More than 10GB of the RAM was deduplicated throughout the whole test (and it had mixed randomized workloads with several apps).
How busy was the small little Celerra?
Interesting indeed.
BTW - today we also published a whole slew of additional Joint Reference Architectures to Powerlink. If you're an EMC customer, a partner, VMware, or anyone looking for information on this - head on over.
BTW - if you look at how pervasive VMware is in the document subjects, you get an inkling of how focused EMC is to make sure we are adapting how our customers want us to - and to be focused on the VMware Infrastructure use cases.
- VDI:
- New EMC Infrastructure for Deploying VDI solution (the one I mentioned above) —This solution is based on Celerra NS20 and demonstrates how to use EMC snapshots in a VMware VDI environment.
- Applied Technology technical white paper about RSA SecurID integration in virtual desktop environments.
- Exchange 2007
- A documented end-to-end messaging solution in a virtualized platform for Microsoft Exchange 2007, leveraging “best-of-breed” technology from Cisco, EMC, and VMware. This helps customers considering an Exchange 2007 deployment to capitalize on the benefits of a complete integration solution from leaders in IT solutions and services.
- New EMC solution for metropolitan Exchange recovery in a VMware environment:
- Features highly available data protection with zero data loss between the primary and secondary site solution, with custom recommendations to help customers with disaster recovery capabilities of EMC MirrorView/Synchronous in a virtualized Microsoft Exchange 2007 environment. EMC CLARiiON and VMware are used for storage and server configuration.
- Building-block guidelines for virtualizing an Exchange 2007 mailbox server role using a real-world deployment scenario. Features validation of CLARiiON
CX3-80 and VMware ESX Server 3.5 hosting the Exchange Server 2007. - New EMC Solutions for Microsoft Exchange 2007 with EMC CLARiiON CX4-480:
- Testing completed to the Microsoft Exchange Solutions Review Program (ESRP) standards. Features performance measurements for Exchange 2007 on the CLARiiON CX4-480 for up to 42,000 Exchange 2007 mailboxes
- New EMC Solutions for Microsoft Exchange 2007 for midsize enterprises include:
- EMC solution for archiving e-mail using EMC EmailXtender with Unicode support—safely archives e-mail:
- Features a demonstration of how to use EmailXtender and DiskXtender to safely archive e-mail to EMC Centera to reduce customer’s user mailbox size, server database footprint, and backup and restore times.
- EMC solution for e-mail backup with Mozy - a Backup with software as a service (SaaS) solution features an illustration of how to use Mozy software to safely back up Microsoft Exchange 2007 Smart Plug-In (SPI) server databases in a virtualized VMware environment using a CLARiiON AX4-5i array.
- EMC solution for e-mail backup with EMC NetWorker: centralized backup and recovery solution features validation of how to use EMC NetWorker to safely back up Microsoft Exchange 2007 server databases to disk storage in a virtualized VMware environment.
- EMC solution for e-mail backup with EMC Retrospect: Secure backup and restore solution describes how to use EMC Retrospect to safely back up Microsoft 2007 server databases in a virtualized VMware environment using a CLARiiON AX4-5i array.
- EMC Solution for Microsoft Exchange 2007 with CLARiiON CX4-480:
- Testing completed to the Microsoft ESRP standards. Features performance measurements for virtualized Exchange 2007 on the CLARiiON CX4-480 up to 4,000 Exchange 2007 mailboxes.
- Microsoft SQL Server - expanded EMC Solutions for Microsoft SQL Server 2005 for midsize enterprises—supported by two new and two updated Reference Architectures. Features:
- Expanded consolidation capability enabled by CLARiiON CX3-40 and VMware ESX Server on Windows 2008 in a Microsoft Cluster Server (MSCS) cluster.
- Expanded protection capabilities enabled by CLARiiON CX3-40, VMware ESX Server, EMC RecoverPoint, and EMC Replication Manager on Windows 2003 and 2008 in an MSCS.
- Expanded test and development, backup and protection, and virtualization capabilities enabled by EMC Celerra NS20, EMC NetWorker, Replication Manager, VMware ESX Server, and Microsoft Hyper-V.
- Microsoft Office SharePoint Server
- Expanded EMC Solution for Microsoft Office SharePoint Server for midsize enterprises, supported by updated Reference Architectures. Features expanded performance, scalability, and backup capabilities enabled by Celerra NS40 and the NetWorker Module for Microsoft SQL Server (NMM 2.0).
- Oracle
- New EMC Scalable Architecture for Oracle Data Warehouse solution—Provides customers with a recommended Reference Architecture and best practices for implementing a flexible, scalable Oracle data warehouse infrastructure leveraging the power of Symmetrix DMX-4 to achieve optimum performance, scalability, and flexibility.
- Expanded EMC Replication Accelerator for Oracle E-Business Suite solution—Provides a validated solution and best practices for cloning the entire Oracle
E-Business Suite (Release 12) environment. It focuses on creating a well-performing storage design for an Oracle E-Business Suite application infrastructure running on a Symmetrix DMX-4. The solution leverages EMC replication technologies—including EMC Replication Manager with Oracle RapidClone utility—to enable high-speed replication with minimal impact to production. - New EMC Scalable Architecture for Virtualized Oracle Applications Innovated Solution—This Innovated Solution is supported by a technical demonstration, technical white paper, and Sales Kit. Leveraging work done at the Santa Clara, CA Innovation lab, this Innovated Solution demonstrates the capability of deploying a virtualized Oracle E-Business Suite test and development environment and also the use of RecoverPoint to failover to a virtualized disaster recovery environment to support continuity requirements.
- New Proven Solutions for Oracle Database for midsize enterprises based on EMC blended NFS and Fibre Channel protocol (FCP) platforms and the Celerra “pure” NFS solutions on VMware ESX Server. Includes:
- Expanded backup capability leveraging Oracle RMAN for virtualized Oracle 10g and 11g database environments
- Expanded protect and failover capability leveraging EMC replication (SnapView, MirrorView) for virtualized Oracle 10g and 11g database environments
- Expanded high availability capability leveraging VMware VMotion
- New protect solution capability for Celerra NS “F” series expanded to include support for Celerra SnapSure writeable checkpoints
Good job! Thumbs up.
Posted by: exchange hosting | March 22, 2009 at 05:34 PM
I visited this blog first time and found it very interesting and informative.. Keep up the good work thanks..
Posted by: Van Sales | July 04, 2009 at 02:50 AM