editor’s note: we asked Timothy Chou to share a preview of one of the themes he might develop in his part of the program next week. This should give us plenty to talk about…
You might find this odd for me to say, but having lived thru the last major shift in infrastructure we might have a few lessons to learn...
The last major move in compute & storage infrastructure was client-server computing. Driven by low cost hardware and software minicomputer and mainframe companies (e.g. Digital Equipment Corporation, IBM AS/400,) were passed up by the rise of client desktops driven by companies like Dell, Microsoft and Compaq. On the server side companies like Sun, Oracle, EMC, Veritas (the Four Horseman of the Internet) grew and ultimately became what you today think of as the traditional enterprise infrastructure.
So why did that happen? And how did it happen? First, it was driven by economics and NOT technology. IBM’s database, operating systems and hardware were technologically superior. I still remember people saying, “We can’t move to Unix, it’s not scaleable enough”, or “We can’t use Oracle, it’s not reliable enough. So why would a technologically superior product lose? Economics. I hear the same thing today, “You can’t move to the cloud, it’s not secure enough”, “You can’t move to the cloud it’s not scaleable enough”. And yet all of the growth in compute & storage is happening by the rapid adoption of cloud computing. Why? Economics.
If you’re a young startup, you can stop reading now, other than giving you a history lesson as Obi-Wan said, “There is nothing to see here, move on”.
On the other hand if you’re a CIO or VP, IT that is managing existing traditional infrastructure and your CEO having gone to the Masters is asking, “So what is our cloud strategy? Then read on.
As with most things, one size does not fit all. So divide your world into your existing and future infrastructure. For your existing compute, storage you should ask “Why move” “What to move” and Where to move?”
So why move? If you look at the fundamental economics of computing the amount of money you spend to manage your compute & storage is at least four times the initial purchase price per year. So in four years you’ve paid 16x the price of the hardware. Why? Because you’re going to spend money on managing the security, performance, availability and change all of the software required to deliver those compute & storage instances. Just consider the number of security patches, testing them and scheduling the roll out as a small part of the true cost of computing. If you move to a compute & storage cloud service the service provider should be standardizing, repeating and automating any of the key availability, security, performance and change management processes. So not only should you get a lower cost (computers are cheaper than people), but also a higher quality service.
And if you want to learn what to move and where to move, stay tuned for the next blog, or come to the kickoff.