I think the following conclusion is correct: Google built a network that is adaptable to the compute requirements of their business.
Bingo. It’s not about the network. It’s about getting the damn thing out of the way.
Google’s Openflow is not your Openflow.
When Urs speaks, datacenter peeps listen.
The problem is we aren’t meeting this challenge. Our infrastructure is broken. Datacenters have the diameter of a microsecond, yet we are still using entire stacks designed for WANs. Real-time requires low and bounded latencies and our stacks can’t provide low latency at scale. We need to fix this problem and towards this end Luiz sets out a research agenda, targeting problems that need to be solved:
* Rethink IO software stack. An OS that makes scheduling decisions 10s of msecs is incompatible with IO devices that response in microseconds.
* Revisit operating systems scheduling.
* Rethink threading models.
* Re-read 1990’s fast messaging papers.
* Make IO design a higher priority. Not just NICs and RDMA, consider CPU design and memory systems.
Google takes another potential step in becoming an even bigger force in communications, LTE, and more. Avaya bid $475 and then bought Nortel enterprise coms for $900m after outbidding Siemens. I suspect there may be a few interested parties bidding on the patent portfolio - Cisco, RIM, Nokia, Ericsson, heck, even Apple and Microsoft could end up at the party. Fascinating.
Google: The Vertically Integrated Content Service Provider?
This project will include:
* an application bundle including a server and web client supporting real-time collaboration using the same structured conversations as the Google Wave system
* a fast and fully-featured wave panel in the web client with complete support for threaded conversations
* a persistent wave store and search implementation for the server (building on contributed patches to implement a MongoDB store)
* refinements to the client-server protocols
* gadget, robot and data API support
* support for importing wave data from wave.google.com
* the ability to federate across other Wave in a Box instances, with some additional configuration
This project will not have the full functionality of Google Wave as you know it today. However, we intend to give developers and enterprising users an opportunity to run wave servers and host waves on their own hardware.
XMPP lives on!