Hi, On Mon, May 05, 2014 at 07:02:44PM +0100, Alan Buxey wrote: > Obviously no love here for VSS etc > > But how is any of this any different not only to other virtual technologies > (be they VLAN, MPLS, OTV etc) but to the code that you all rely on from cisco > for the other things that keep the network running (spanning tree, EIGRP, > OSPF, FIBs etc) ? > > Surely if you have code/quality issues then trust in ANY of the stuff is an > issue and this isn't just a knee jerk reaction in classic luddite fashion?
Of course we don't trust any code Cisco produces. The interesting question is, what effects will a "oh, this packet looks nasty, crash!" bug have? In our setup, we use standalone boxes (no VSS), with different IOS *trains* on them, like 12.2SXI and 15.1SY, or such. With the hope that certain classes of bugs will not affect both trains in the same way, so one box might crash while the other will continue to work. Adding complexity to the control plane while at the same time coupling things more tightly increases the likelyhood for "nasty bugs", while at the same time increasing the *effect* of a nasty bug - double no-go in my book. And yes, I'm not drinking the "you can run your whole network using one of our nice SDN controller boxes!!!" kool aid. gert -- USENET is *not* the non-clickable part of WWW! //www.muc.de/~gert/ Gert Doering - Munich, Germany g...@greenie.muc.de fax: +49-89-35655025 g...@net.informatik.tu-muenchen.de
pgpkIgzonzWzK.pgp
Description: PGP signature
_______________________________________________ cisco-nsp mailing list cisco-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/