As the traditional IT vendors try to build infrastructure stacks and package them at a premium cost; there is part of me which is beginning to wonder whether this makes as much sense as it appears. Obviously it makes sense for the vendor who is trying maximum their profits but for customers?
I read tweet after tweet from various of the stack vendors about how they are winning service provider business but does this show a remarkable lack of imagination and technical chops from the service providers? Is it time for more service providers and larger enterprises who are in the market for many thousands and many tens of thousands of servers to look at cutting out the middle man and working with the ODM’s directly.
Obviously we already have companies like Google and Facebook who are already working with the ODMs; the ODMs are hence gaining experience in dealing with end-customers and dealing with a more application-centric view and might this not start to become a threat to the more traditional vendors? And might we see them start to work with companies which have smaller scale requirements, large service providers of all kinds might find the option of servers which fit their requirements exactly very attractive.
HP for example are offering what is basically a data-centre in a shipping container but might this be something that the ODMs might be better placed to build and provision for the customer who has such a requirement.
And if big data is a real thing; ODMs might well be very well placed to attack this market? White-box storage appliances utilising open-source software such as Ceph, Lustre and Gluster for example? Networking appliances built on commodity hardware?
Now, I don’t think this is an overnight change in the market and it’s probably not an immediate threat to the traditional vendors but I can see an interesting tension between the ODMs and the traditional vendors developing.
As an end-user of IT; it’s all good, choice is good both in vendors and architectural principles.