On 3 June 2016 at 18:39, James Bensley <[email protected]> wrote: >> Not all platforms use TCAMs. Lot of Juniper kit, like MX, QFX10k, PTX >> use various types of DRAM solution, this makes FIB usually not your >> bottleneck, search time to larger database becomes an issue too. > > If vendors switched to using hashed tables we would get consistent > look up times irrelevant of table size.
Vendors are using all kind of algorithms and inventing new ones. It is by far more complex than 'trie' or 'hash', particularly when using DRAM memories, like Juniper is. I believe Juniper has done lot of academic work on this, maybe partly procket legacy. But I think it's trade-secret not published work. -- ++ytti _______________________________________________ cisco-nsp mailing list [email protected] https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/
