At 1:40 AM +0000 8/3/03, " Chuck Whose Road is Ever Shorter " wrote: >""Howard C. Berkowitz"" wrote in message >news:[EMAIL PROTECTED] >> At 3:35 PM +0000 8/2/03, Charles Cthulhu Riley wrote: >> >Less IP addresses used? >> >> Typically, the advantage of P2P is that you can impose individual >> policies on each spoke. A basic such example would be bandwidth >> matching the CIR if all CIR's are not the same. Spoke-specific >> access lists would be another. Routing configuration generally is >> easier. >> >> You also get finer granularity for SNMP, accounting, etc. >> >> P2M might slightly conserve IP addresses, but, more significantly, it >> conserves Interface Descriptor Blocks (IDB) and interface buffers in >> the IOS. In some respects, it's more intuitive, although the routing >> configuration is more complex. > > >This was probably an important issue several IOS versions ago. These days, >with limits in the thousands ( maybe up to 10,000? ) descriptor blocks >available, even on the lowly 2501, this is no longer an issue. > >As I once said in another lifetime, changes in hardware and software have >led to less concern with traditional design issues that were centered around >scarce resources.
Agreed that the IDB limit is not the issue with appropriate releases. When you consider interface buffers are allocated to each subinterface, however, that's a different memory impact on a small router. Admittedly, that isn't as major with the newer platforms. On a 2501 with 2MB shared RAM (where the buffers go), it's major. > >If I can trust the Cisco writings on the topic, trhe more modern QoS >mechanisms have even led to more effective use of WAN bandwidth, which has >continued to be the real bottleneck in networking. That bottleneck, as you know, isn't necessarily absolute bandwidth, but queueing delay in access to bandwidth by latency-sensitive applications. >Tools such as RED, WRED, >and tail drop have helped alleviate the problems associated with the >phenomenon of global synchronization. I suspect the work of the IETF and >queueing theory researchers over the past decade of so have led to a more >effective use of bandwidth, meaning that more data can use the same link. If >I understand correctly. Message Posted at: http://www.groupstudy.com/form/read.php?f=7&i=73431&t=73431 -------------------------------------------------- **Please support GroupStudy by purchasing from the GroupStudy Store: http://shop.groupstudy.com FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html