Atmos becomes the next Invista. Okay, my last blog mentioning EMC generated some heated comments, and there is no doubt that this blog may generate even more. For those that sell Atmos, you have done great work to develop and bring new technology to a market that has been hungry for cloud products. For those that have deployed Atmos, I am sure you have the full support of EMC behind you and that any product mix changes by the vendor will be integrated into your overall strategy.
However, I have spoken to a number of distributors and IT customers who have gone through both the paper analysis and actual proof-of-concept testing of Atmos. Given the choices available in the market today, nearly all of the IT customers chose to deploy other storage solutions. Why? Simple answer: They have a requirement to scale past 100 million objects. The customer that chose Atmos had no such object number requirement, but did have applications built to use Atmos' type of interface.
The distributors I have spoken with tell me that it is their opinion that "EMC will pull Atmos from the channel and move to isolated sales via their direct sales channel." I am told it is because of the error codes it spits out once the limited object store goes beyond 100 million objects. The distributors informed me that it was "becoming increasing more difficult to position the product within environments that need massive scale-out capability."
100 million objects may sound like a lot. In actuality it is not much given today's massive data storage requirements. This is especially true when you consider that a single Web-based company--even a single department of an entertainment company, for example--can have over 400 million JPEGs that need to be stored in the cloud. Given the recent acquisition of Isilon by EMC, it's my opinion that EMC will move Atmos' development into the Isilon team and integrate Atmos' object store within the Isilon Clustered File System.