30 drives in a 3 U platform. Includes servers, DAS SAS or SATA, IP connectivity between servers and IP connectivity to client network. Sounds like a HPC cluster in-a-box?
Configuration details:
- 300 TBs in a 44U rack
- 10 - SAS / SATA disk shelves
- 12 1 U servers
- Dual Switches
- 24 gigE server ports
- One 10 gigE port for uplink
- Front-to-back cooling
- Variable server to drive ratio (1:30) defined above
- Config for capacity or performance
- Upgrade CPU and storage independently
- HA config options available
- Simplified serviceability
Note: Not CX DAEs (visibly they look the same to me, but apparently they are native SAS shelves) but they have the same footprint. The equivalent of 2 DAEs are mounted in the same 3 U on a moveable shelf. Essentially 2 DAEs in the same 3 U vertical footprint.
InfiniFlex is customer serviceable, customer can spare all parts or the customer can opt for a standard EMC support contract.
InfiniFlex 10000 Server Specs:
- 64-bit dual-core
- 4 GB of RAM
- Single Power Supply
- CD / DVD-RW (boot image)
- Standard 19" NEMA rackable
- 20" deep
- Adaptec 3085 disk controller
Use cases:
- Web 2.0 deployment model
- Proprietary applications
- Availability by replication
- Preference for self-service
- Web 2.0 use cases
- ISP services
- Social networks
- Cloud computing
- etc…
- Content depots
- Documents
- Videos
- Surveillance
- Research repositories
Should have called this session Site Recovery Manager brainwashing revisited.
I find it interesting that we are still spending time talking about Replistor and MirrorView as viable Exchange replication technologies. I thought we determined years ago that they just don’t work.
It would make more sense to me when talking about Exchange 2007 and replication to spend more time on Local Continuous Replication (LCR), Cluster Continuous Replication (CCR) and Single Copy Clusters (SCC). With applications like Exchange enterprises will more to adopt the technology with the tightest application integration (e.g. - Oracle RAC). I would have preferred a discussion focused on application integration.
Just my 2 cents.
I have been really busy the past few weeks so the blog has slowed to a snail crawl, although I do have an interesting blog that compares some of the more popular Windows data migration tools so stay tuned for that. This week I will try to report live from EMC World and the sessions that I attend, note that I will be listening, typing and publishing so please excuse spelling and grammar mistakes as I will not be spending time proofreading.
I am now sitting in a session entitled “Simplifying Disaster Recovery with VMware and Celerra Replicator” This session should be entitled “VMware Site Recovery Manager and Celerra Replicator”
Since I have been typing this blog I may have missed a few details but here is what i have heard thus far:
Site Recovery Manager (SRM) is supported in the following environments:
- >= VMware 3.0.2 and Virtual Center 2.5
- >= DART 5.6
- iSCSI Only (No NFS Support)
One topic that I feel the need to address is the thought that SRM replaces Disaster Recovery (DR) Runbooks. I could not disagree more, under ideal conditions SRM automates the tasks traditionally documented in the DR Runbook but it is important to remember that SRM is another piece of software that is added to an already complex mountain of software and process, I do not believe this is a replacement for the DR Runbook but rather another infrastructure component that needs to be documented as part of the DR process.
Note: VMware will only support RDM devices as beta? What the heck does that mean?
SRM only support crash consistent copies, essentially no support for application awareness and Replication Manager.
I asked the question with EMC and the application vendors (Microsoft) pushing for application (Exchange, SQL) awareness via VSS and other application aware APIs what is the plan for SRM and application awareness? The answer I received nebulous at best, to paraphrase the presenter I think he stated that you would need to determine if Exchange is core to your business; last time I checked Email is a fairly critical component of doing business in 2008.
Needless to say to this point SRM is not making a good case for jettisoning the DR Runbook.
Talk to you from Tucci’s Keynote next.
