Large Exchange mailboxes, Low cost disks – IX12 makes it happen
The idea of large mailboxes with low costs disks is something Microsoft loves to promote with Exchange 2010. We are all well aware of the I/O improvements with ESE in Exchange 2010 and the use of large SATA disks has made the large mailboxes a reality for customer who prefer SATA over FC or SAS disks for their cost efficiency and growing capacity.
In May of this year, EMC’s Iomega division released the StorCenter IX12-300r Network Storage Array which is aimed at small to medium business customers. The base configuration starts at 4TB and can be expanded up to 24TB in a single array in either a file or block level (iSCSI) configuration.
Due to the low cost, but large capacity of the IX12 we decided that these would be a great platform to show how well Exchange 2010 can work with low cost storage enabling large mailboxes that customers want. The initial result was our Exchange Solutions Reviewed Program (ESRP) submission for 260 users on the IX12 where we showed a building block for 260 users with 10GB mailboxes in a 3 HA copy Database Avaliability Group. Yes, even small to medium customers can have large mailboxes and HA/DR!
The ESRP for the IX12 was a great start, but we wanted to take it further into a more capable configuration up to 1000 users and show how we could push the IX12 farther and showcase how several IX12′s could be put together in a Site Resilient configuration with large 4GB mailboxes and in an Exchange Native Data Protection scenario (i.e. backupless) with Lagged copies for point in time protection. Oh yeah, we also wanted to show how Exchange 2010 would be virtualized in a low cost Microsoft Windows 2008 R2 Hyper-V environment for customers who have made the decision to virtualize Exchange with Hyper-V.
The result is contained in a new whitepaper that we just released called “EMC Business Continuity for Exchange 2010 Enabled by EMC Iomega IX12-300r and Microsoft Database Avaliability Groups”
The physical environment that we designed in this Solution looks like this:
Some of the key proof points contained in this whitepaper are how both HA and Lagged copies can be distributed across mailbox roles and how the Site Resiliency process works with DAC enabled.
This is something that is commonly misunderstood and we show you how we did the testing to simulate a site failure. The other thing we captured is what we saw with DAG seeding performance and this testing from both and local and remote scenario.
For instance, we saw in a local failure scenario that a user was back online after a single server failure in 21 seconds. Jetstress performance for this array was also very good as achieved transactional IOPS/sec were around the 76-80 per server mark with read and write latency between 15-17ms.
Overall, we were very pleased with the results from this Solution and the findings show that even with a low cost array like the StorCenter IX12 with a powerful hypervisor like Microsoft Hyper-V, it can be a very cost effective and powerful solution for Exchange 2010 to enable a large mailbox experience with HA and DR capability, even for the small to medium sized customer. The StorCenter comes in about 4-5k retail, so check with your favorite reseller on these.
If you want to read more, please see the detailed whitepaper at: http://www.emc.com/collateral/hardware/white-papers/h8035-business-continuity-exchange-iomega-wp.pdf
Until next time,