With Coho Data:
1. Customers can buy their own compute from Cisco, Dell, HP, Supermicro, etc.more economically
2. If you only need storage capacity, why be forced to also have to buy more compute? Therefore it makes sense to keep them separately from an economical standpoint
3. When the CPU is in the storage it competes for performance cycles, when it's separate it doesn't
4. When you only need storage capacity and are forced to buy CPU, in a VMware environment, you are forced to buy more server licenses
5. So in a semi-converged approach you get the benefit of buying your compute as needed for your mainstay applications and also use our CPU in background for Microservices, avoiding the cost of having to invest in additional compute for Hadoop, Transcoding, Splunk, and other apps...so you get the best of both worlds! Not to mention all of the other cool stuff that Coho is delivering today, like linear scale-out from 7TB usable to petabytes with all of the performance you need to support your virtualized environments today, and in the next several quarters, your physical environment as well from a blend of all flash arrays featuring all NVMe PCIe and SAS SSD flash, to our original hybrid arrays (3 models) with NVMe PCIe flash and HDD. Coho still believes that HDD makes sense today for all of your cold data, but agrees that at some time in the future the economics will lean towards putting in some layer of cold flash, but IMHO that day is not here yet. Since our system is extensible, when that day is a reality, we simply blend that new cold flash tier into today's cluster easily and automatically. If you haven't checked out Coho Data yet, give us a call so we can share what the co-creator of the XEN hypervisor who still consults with Citrix and his team are up to in solving the most difficult problems in the storage world; namely reducing complexity, eliminating islands of storage, while lowering CAPEX and OPEX costs by 50%!