Last week I attended the Cloud2020 Summit in Las Vegas and came away with an entirely new perspective on Cloud Computing that I wanted to share. Before I do that, a quick note on the event; all sessions were videotaped and should be available on the event site shortly. Also, thanks to Ben Kepes and Krishnan Subramanian for putting together such an exclusive group of speakers and Mark Thiele for hosting us at the Switch site in Las Vegas.
The event spent the morning focusing on the most overlooked portions of cloud computing; the physical data center and the storage tools behind the scenes. Most people think of the cloud stack as simply the server, hypervisor, virtual machine/operating system, and secure connection or access via the web to that virtual machine. You could even argue that most people take a complete software view of the cloud stack and simply think about the hypervisor, virtual machine, and management layer; see CloudStack, OpenStack, vCloud, etc. This event opened my “eyes” to the substantial impact that a physical datacenter and storage solution offer to the cloud.
The physical datacenter is really the driving force behind cloud computing to realize its full potential. Without the ability to have a secure facility to locate massive amounts of computing power and storage at a reasonable cost and have the power redundancy to ensure 99.9999% uptime, cloud computing would simply not be possible. The complexity and innovation at the Switch datacenter was breathtaking as that single datacenter was more efficient in power usage, temperature control, and computing power availability than 3 typical datacenters. Thus, the Switch datacenter alone provided economies of scale for environmental factors and raw computing than could possibly have been achieved with all the users of that site running their equipment independently.
The second focus of the morning was on the changes coming in cloud storage. The consensus at the event was the complete dominance coming to cloud storage via solid state devices with the exception of long term Write Once Read Never (WORN) storage. More and more information needs to be stored for the long term with little to no chance of ever being used again; thus the return of tape. The return of tape storage was not something I considered and was surprised at the potential size of this market. In fact, I have begun looking into companies making tape storage devices as well as the tape itself as potential long term investments (NOTE: I am not a financial advisor and you probably don’t want to invest on my ideas).
As for the rest of the day, there were some excellent debates about who will dominate cloud platforms in the near future and the impact of the various technologies and players. There are several other blogs that have focused on these points so I will leave that for you to spend more time on. Finally, there was an excellent talk by Joe Weinman, author of Cloudonomics that I highly recommend viewing. Ben and Krishnan are planning to host this event again next year and I cannot emphasize enough the high quality of the event, speakers and participants.