by Pablo Carlier.
Getting Big Blue into the private cloud world:
The acquisition will give IBM a significantly larger toolbox for tackling high-performance and technical computing applications such as “big data” analytics, simulation, and product design. Platform Computing also brings along technology that will help round out IBM’s cloud computing offerings.
From Ivan Pepelnjak’s blog, comes a nice, yet daunting thought: the Holy Grail of modern networking, massively-scalable multi service Layer 2 domains, has the ultimate goal of allowing MAC-to-MAC visibility between servers, mainly for virtualization purposes.
… everyone is running to join the gold rush, from Cisco’s FabricPath and Brocade’s VCS to HP’s IRF and Juniper’s upcoming QFabric. As always, the standardization bodies are following the industry with a large buffet of standards to choose from: TRILL, 802.1ag (SPB), 802.1Qbg (EVB) and 802.1bh (Port extenders).
The only viable argument the whole industry has for the push toward large(r) layer-2 domains is VM mobility – if you want to migrate a live virtual machine across the data center and retain its sessions, it has to stay in the same VLAN…
Ivan’s argument is surprisingly simple: it would not take much for a virtualization vendor to wake up, smell the coffee, and enable an IP-to-IP VM mobility mechanism. His conclusion is indeed clear:
All of a sudden, all we need in the Data Center is layer-3 connectivity designed and implemented using the same mechanisms we’ve been using for the last 30 years to build the (somewhat scalable) Internet. The only reason for the layer-2 gold rush is gone.
I believe that VM mobility is not the only reason behind these efforts though. IP-based mobility would not by itself solve the problem of providing a full-fledged network interface to the VM. It would not solve inter-VM policy making without further enhancements. It would not solve the management overload associated to network device sprawl.
But still, it would mean a paradigm shift for innovation in the Data Center space.
Yesterday I wrote about Intel going networking. Now it’s Dell that’s going after Force10 and their Zettascale line. It’s gonna get funny in the networking space.
Huge play by Intel with this announcement. Fulcrum manufactures switching chipsets, effectively turning Intel into a networking vendor for the Data Center space. Huge.
Via Rank & File and After Hours I found this insightful piece by Signal vs. Noise’s David about consumerization of IT. I find his points very interesting, especially how he believes IT departments have not caught up with the speed of tech expansion:
It’s the same forces and mechanics that slowly turned unions from a force of progress (proper working conditions for all!) to a force of stagnation (only Jack can move the conference chairs, Joe is the only guy who can fix the microphone).
IT consumerization is here to stay. But I think IT needs to change from being the “policy enforcer” to “the one that makes it happen”.
Users used to passively receive tools to do their jobs, training to understand those tools, and support when those tools failed. Now they actively demand access to the technology they already use, love and know in and out at their work place, and they look to IT to make that happen.
Martin Casado from Network Heresy continues his series on software switching and virtualization with his latest article.
There are very valid points in his post about what the real costs in terms of latency, CPU resources, throughput, and others. His summary: software-based hypervisor switching does not have that much of an impact on these as one might believe, and, in his own words, “kicks mucho ass”.
I think Martin is missing one point, though: complexity. That’s where I believe passthrough “VM-Fex” solutions really help. Keep reading to find out how.
