Posted by: Denny Cherry
EMC World, EMC World 2010
So yesterday was day 1 of EMC World. I attended some great sessions (and one not so great one).
The first session that I hit was the futures for the EMC CLARiiON’s FLARE software. For those that don’t know what FLARE is, FLARE is the software which runs the array and handles all the functions of the Array. The next release will be FLARE v30. If you are a CX3 or older customer this new release will be of no use to use as this version only supports the CX4 array.
Some of the new features which are being included are a totally new management interface called Unisphere. This new interface will give a single interface for your EMC CX arrays as well as your Celera devices and RecoverPoint. Eventually other EMC products will be integrated into Unisphere with products such as Replication Manager coming hopefully in 2011.
Now don’t worry, you don’t need to go out and buy all new arrays to take advantage of Unisphere. Unisphere will work with existing arrays back to FLARE 19 for the CX arrays and DART 6.0 for the Celerras which should cover most of the arrays out there.
Unisphere will give you a single management interface, replacing the slow NaviSphere tree that we are used to seeing with a tabbed interface. However once you get into the actual tasks the screens will look very similar to what we see today. Now for those that can’t upgrade to FLARE 30, or you aren’t ready to upgrade to FLARE 30, there’s an off array version of Unisphere that you can install on a Windows server and use that as your management endpoint to get the benefits of Unisphere without the pain of upgrading your array.
Now sadly Unisphere is still Java based, but it is a total rewrite, and with some tongue and cheek answering we were told that not all the performance problems of the current NaviSphere were because of Java. To me this says “Some of our code blows”. It was probably as honest as an answer you can get with marketing people wandering around the building.
Something which is a little more exciting than a new UI is the new thick pool of LUNs. When you create a LUN pool you’ll be able to create the LUN as a thin LUN or a thick LUN. A thin LUN is just like it is today, the LUN is created as 2 Gigs then new space is allocated in 1 Gig chunks. With a thick LUN the LUN is created from the disks in the pool as it’s full size. Now where this comes into play is with the new FAST technology will monitor the 1 Gig chunks of the disk, and move them around within the array so that hot chunks are moved to faster storage, and cold chunks are moved to slower storage. This is all done within the same pool.
A single pool of disks can have both thin and thick LUNs in it, and it can have disks of different sizes and speeds. When you create a pool you aren’t limited to the 16 disk limit that you have for a RAID Group today.
Another nice thing about the LUN pool is that you can expand the LUNs without having to create meta-luns which reduces your overall LUN count allowing you to create more LUNs.
Now with pooled LUNs (both thick and thin) the 1 Gig chunks are made up of 8k blocks. Within the 1 Gig chunk the 8k blocks are contiguous, but the 1 Gig chunks won’t be contiguous for a single LUN. So you will have some fragmentation issues to deal with.
The last big thing that I learned about yesterday is that with FLARE 30, the CX4s will be able to use the Enterprise Flash Disks (EFDs aka SSDs) as cache. You can take a RAID 10 of EFDs (only RAID 1 or RAID 10 is supported) and mark the RAID Group as Flash Cache. This will allow you to have up to 2 TB of additional cache on the array which can be used as a read or read/write cache. Now this RAID Group will be dedicated as a Flash Cache RAID Group so you can’t create LUNs on this RAID Group. This Flash Cache technology works best when used on very bursty work loads. Typically it only takes 2-4 disks to create a really good, large Flash Cache.
Now the downside to Flash Cache, is the cost. These EFDs cost about $20k each from EMC, so that’s $40k-$80k in EFDs from EMC. During the session the speaker said that this was supposed to be a less expensive was to increase your array’s cache without downing the array and increasing the DRAM. Now I’m not rocket scientist, but if the disks are 400 Gigs each at $20k, and I put in two for Flash Cache, 800 Gigs of RAM (400 Gigs for each storage processor) will be a lot less expensive than the $40k for the flash drives. Now granted you won’t be able to get 400 Gigs of RAM into a single SP, but I think that EMC Marketing might want to come up with a better message than this.
Hopefully you like some of these new features, and I’ll see you tomorrow.