Blog Posts in December 2012
When I’m asking the yearly subscribers whether they’d like to renew their subscription, I promise them new content every 2-3 months (4-6 new sessions per year). 2012 was definitely a good year in that respect.
It started with the access network part of large-scale IPv6 design and deployment webinar, then there were two Data Center Fabrics update sessions (in May and November), scalability part of the cloud computing networking webinar, and a DMVPN design session.
12 months and ~210 blog posts later, it’s time for yet another “That’s It” blog post. Another exciting year has swooshed by, and I’d like to thank you all for the insightful comments you made, the great questions you asked, and the wonderful challenges you keep sending me.
If at all possible, now’s the time to start shutting down the pagers and smartphones, and enjoy the simpler (and less stressful) life with the loved ones. Have a great holiday season and all the best in the coming year! I’m going offline ... right now ;)
In August 2011, when NVGRE draft appeared mere days after VXLAN was launched, I dismissed it as “more of the same, different encapsulation, vague control plane”. Boy was I wrong … and pleasantly surprised when I figured out one of the major virtualization vendors actually did the right thing.
TL;DR Summary: Hyper-V Network Virtualization is a layer-3 virtual networking solution with centralized (orchestration system based) control plane. Its scaling properties are thus way better than VXLAN’s (or Nicira’s … unless they implemented L3 forwarding since the last time we spoke).
Justin asked an interesting question in a comment to my IPv6 On-Link Determination post: do we need HSRP for IPv6 as the routers already send out RA messages? Pavel quickly pointed out that my friend @packetlife already wrote about it, concluding that you could use RAs unless you need deterministic sub-second failover.
However, there are (as always) a few more gotchas:
A year ago I asked you to help me plan this year’s webinars. Let’s see how well I matched those expectations:
During the IPv6 Security webinar Eric Vyncke described various mechanisms you can use to implement RA Guard on Cisco IOS, including private VLANs, port ACLs, RA Guard Lite and full-blown RA Guard available in recent Cisco IOS releases.
When testing the OSPF graceful shutdown feature, I've encountered an interesting OSPF feature: if you force a change in LAN DR router (other than rebooting the current DR), you'll end up with two network LSAs describing the same LAN.
This blog has been sitting in my Draft folder for years, so Cisco IOS behavior might have changed in the meantime, or it might have been a transient and/or race condition. Nonetheless, I still find it interesting.
November 2009 was the month of Hierarchical Queuing Framework and other QoS goodies, and the month I started my periodic “we’re missing IPv6 in the Data Center” rants (three years later, IPv6 is still MIA in virtual networks).
In November 2007 I was writing about BGP default route ... and I’m still getting questions about it.
The second scenario Brad Hedlund described in the Clos Fabrics Explained webinar is a large leaf-and-spine fabric using 10GE uplinks and QSFP+ breakout cables between leaf and spine switches (thus increasing the number of spine switches to 16).
Ronald sent me an interesting question: What's the point of having a secondary path set up for a certain LSP, when this LSP also has fast-reroute enabled (for example, with the Junos fast-reroute command)?
The idea of having a pre-established secondary LSP backing up a traffic engineering tunnel was commonly discussed before FRR was widely adopted, but should have quietly faded away by now.
A while ago I wrote a blog post about remote ND attacks, which included the idea of having /120 prefixes on server LANs. As it turns out, it was a bad idea, and as nosx pointed out in his comment: “there is quite a long list of caveats in all vendor camps regarding hardware in the last 6-8 years that has some potentially painful hardware issues regarding prefix length. Classic issues include ACL construction and TCAM specificity.”
One would hope that the newly-release data center switches fare better. Fat chance!