EMC acquires ScaleIO, but why?

EMC's ScaleIO Acquisition

I have been thinking about EMC's acquisition of ScaleIO since it was publicly announced yesterday and thought I would share them with you fine Internet readers.

What is ScaleIO

Before we start to talk about why EMC would buy this company, lets talk about what they do. I have a conversation with them about 2 months ago when I first heard about them. The long and short of it is that they take the storage that is available on machines and make it available as a distributed storage pool. The selling point here is that it is a software only solution. You bring your own storage and with the software you can have a SAN without all the expensive hardware from a large vendor like EMC or HP. This would in turn lead to lower costs and this is what they focus on in their marketing. 

The software can be used on enterprise linux distributions and VMware ESXi. The ESXi piece is what got my attention and which is why I gave them a call to talk about the solution. I was thinking about using this in a VDI install. The idea would be get some powerful standalone machines with a decent amount of local spinning disks and SSDs for caching. With this, I would have an island of VDI compute and storage that could scale out both storage and compute as needed. This could keep costs down and still be flexible when needed. Seemed like an interesting idea. At first that is...

But what about vSAN?

Then I remembered that VMware has a product in waiting that solves this problem in a more native way. It is called vSAN. The ideas behind vSAN and ScaleIO are the same. Local disk in each ESXi host that is available as a storage pool for all ESXi hosts to use. The difference is that it would be native to ESXi and therefore more easily managed by vCenter and therefore more easily usable by vCloud ( hopefully ).

So why?

So why would EMC, the semi-parent company of VMware, by ScaleIO? At first, I guess they could sell it as software SAN solution to fill in the gaps in the lower end of their portfolio but I don't think this makes sense. That makes more sense for Dell than EMC. They sell compute and storage already.

This solution could be used with something like XtremeSF to make all the cards in all the ESXi hosts available as a distributed flash SAN. Then with XtremeSW, you could use the capacity on all XtremeSF cards in the cluster for caching. But then again, this is what vSAN could provide as well maybe without all the caching at first. Seems like it is competing with itself.

Then I realized how this might be used. Remove ESXi from the previous scenario. If you don't do this within a VMware environment, then ScaleIO starts to makes sense. You still need a distributed filesystem or a way to arbitrate access to the distributed block device. The is quite a curios scenario since EMC, along with VMware, is pushing customers to use ESX and the vSphere suite for all apps. Maybe there is a large enough market out there of people not virtualizing all apps where this might get EMC a foot in the door.

Another usage scenario where this might be a good fit is HPC. The HPC folks tend to use commodity hardware and distributed filesystems for storage in the first place. With ScaleIO, they get a nice solution and it also gives EMC a way to get in the door with HPC shops.

Lastly, maybe EMC acquired this product to remove it from the general market. It is a competing product for what EMC offers in hardware today and also for what VMware is going to offer very soon in software.

What do you think? Leave some comments below.