![]() ![]() Next, existing links can be fattened, or direct connections between backbonerouters can be added as required or as is cost-effective. Building the Backbone Topologyīecause you have a basic need for resilience in the backbone, a good startingpoint for the backbone topology is a ring connecting all distribution networks.This ring could represent the minimum cost of WAN circuits, compromised by aninitial estimate of major traffic flows, and possibly some very particular delayrequirements (although this is rare, with notable exceptions beinghigh-performance networks). Data network requirements analysis is a relatively new art. This initial design can besubsequently refined quite effectively by statistical analysis of traffic levelsafter the backbone is operational, and the availability of new WAN technologiesis known. Most backbone topologies are, therefore, initially designed based onfinancial constraints, such as user population density, or applicationrequirements and WAN service availability. Round-trip time and TCP window sizepermitting, any user can burst traffic up to the full line rate of the trunk.Furthermore, the routing complexity in a full mesh can consume bandwidth,computational, and operational management resources. Chapter 14 explores how to bypass best-effort by providingdifferentiated service in IP networks.Ī full PVC mesh can also obviate one of the benefits of multiplexing, or trunking, in a best-effort network. One good reason is that the resources required by any TCP/IP sessionare not known a priori, and IP networks are traditionally engineered asbest-effort. ![]() Although traffic engineering calculationsand circumventing congestion are common in the telephone network, common IPnetworks and their associated routing protocols do not provide this capabilityas readily. Moreover, PVC sizing requires that the traffic levels between any twodistribution networks should be well understood, or that the network has thecapability to circumvent congestion. ![]() This creates expense in WAN circuitry, as well as inrouter and WAN switch hardware (channelized or ATM technology can reduce theseissues). However, in terms of minimizing hop count within the network, the fullmesh approach has several drawbacks:įirst, given N regional distribution networks, you must have N(N-1)/2backbone links in the core. The second option is to simply connect all distribution networks with a fullmesh. ![]() If thecarrier's up-time guarantees are not sufficient, you have no choice but todesign a backbone that is resilient to link failure. Ultimately, however, any backbone will include WAN links thatrely on a great deal of equipment and environmental stability for theiroperation, which represents a real risk of ultimate failure. First, you can createmore reliable routers through the use of "carrier-class"characteristics, such as multiple CPUs, power supplies, and generators and evenredundant routers. Reliability can be acquired by employing two methods. Given its position at the top of the network hierarchy, tworequirements of the backbone topology are clear: it must be reliable and it mustscale. A nationalinfrastructure usually forms a significant part of the operational cost of thenetwork. The purpose of the backbone is to connect regional distribution networks and,in some instances, to provide connectivity to other peer networks. With the move toward multipleclient-server and peer-to-peer relationships, the choice of core networktopology is not as clear. This did cause the center of the star to become a single point offailure, but because no real traffic flows existed between spokes on the star,this was not a major cause for concern. In early data networking, the topology for the network backbone wasrelatively simple: Operations were centralized, so a star topology made the mostsenseand, in some cases, this was the only topology the technology wouldsupport. ![]()
0 Comments
Leave a Reply. |
Details
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |