Quite interesting read:
HP is trying to hire about anyone who can spell cloud while they’re settling in at the OpenStack project.
But what really caught my attention is that bit
Even OpenStack leader Rackspace is being challenged when it comes to rolling out more services: The Reg has learned that OpenStack is running on just 350 of its hosting servers, more than a year after Rackspace helped unveil the project.
Rackspace is using OpenStack on a whopping 350 servers? Thats like 1% compared to big setups like amazon?
When you compare this to all the millions of marketing dollars, partnerships that are pressed into OpenStack to make it the #1 solution, you see an interesting angle
- Rackspace started this half ready and not even internally using it
- Many jump aboard to join their dev efforts, safe money and “be there”
I guess when they found that OpenStack is not the big killer solution yet they just cranked up the marketing to ensure OSS competitors like OpenNebula are out of the standard “cloud talk”, and probably also scramble to get up something working.
So if HP joins there with a lot of engineering power (and they have spent many years in virtualisation research in their labs already), then they’ll be able to take over OpenStack down-handedly.
HP brings along all hardware and networking options. They have looked into stuff like virtualized networks, have had their own switches for the blades for quite some time, and also have a large networking line. They have multiple storage offerings that fit into the cloud hosting world already, and have been involved in key technologies like Xen very early (i.e. the profiling solution for Xen was added by HP at least 5 years ago. So you could say they know it at its core). This combined feature set would make them an ideal choice for anyone moving upwards from SuperMicro or Dell to the upper tier vendors by itself.
Now with OpenStack membership HP suddenly moved into a position where neither IBM (they love Director and Tivoli too much?) or Sun (they’re dead) or Fujitsu (dynamic datacenter – we have the powerpoint presentation!) can go.
Cisco has been there and UCS is much better gear in a much better solution (IMO) but they failed to
tell people about it give it the same cloud angle, and they can’t provide all parts of the infrastructure.
HP might be able to pull just the same kind of vendor lock in like IBM managed it with its mainframes back in the 70s. Quite disgusting when you think of the incredibly bad web interfaces HP keeps adding to its software.
On the other hand, my own always-pessimistic reading of the El reg article is:
Haha, OpenStack is a hoax just like I was told last year.