In determining the best route to a destination, different routing protocols use a number of different measurements. These measurements are called metrics. Each routing protocol uses one or more metric to calculate the best route to a particular destination. The most common metrics include path length (hop count), reliability, delay, bandwidth, load, and financial cost of a link.
Another major difference between routing protocols is how they handle updating each other with current information. There are many methods of doing this. Given these major differences, routing protocols are broken into two main categories: Distance Vector and Link State.
Distance Vector protocols include RIP and IGRP. They send their entire routing tables out in all directions at regularly scheduled intervals.
Link State protocols are more advanced than distance vector protocols because, unlike distance vector, they do not send periodic routing updates. Link State protocols include OSPF, NLSP, BGP, and IS-IS.
They send partial routing tables (of their own networks) to everyone and then send updates when necessary.
Wednesday, November 04, 2009
Cisco Hierarchical Internetworking Model
Cisco characterizes networks by using a three-level hierarchical model. This helps clarify the purpose of each device in the network. Cisco calls them the Access, Distribution, and Core levels. They are described here:
Access level
Where end (user) nodes access the network. These are closet or desktop routers or switches. If the network is designed well, a good portion of the intra-workgroup traffic will remain at the Access level. VLANs are implemented at this level. Some basic filtering can happen here, but it is usually implemented at the distribution level. If there are WAN connections, they would connect the Access level to the Distribution level.
Distribution level
Aggregates the Access level connections to the Core level. Oversubscription of bandwidth happens here (150 10Mbps access-level users do not require a T1 to the Core, because they will not all use this bandwidth at the same time). Most Access Lists, compression, and encryption should be implemented here. Devices at this level are almost exclusively routers.
Core level
Concentrates all traffic that needs to transverse the network. The focus here is switching traffic as fast as possible. ATM, Gigabit, SONET, and other high-speed technologies are usually implemented here (although it is becoming more common to see higher speeds closer to the Access level, as equipment prices drop).
Access level
Where end (user) nodes access the network. These are closet or desktop routers or switches. If the network is designed well, a good portion of the intra-workgroup traffic will remain at the Access level. VLANs are implemented at this level. Some basic filtering can happen here, but it is usually implemented at the distribution level. If there are WAN connections, they would connect the Access level to the Distribution level.
Distribution level
Aggregates the Access level connections to the Core level. Oversubscription of bandwidth happens here (150 10Mbps access-level users do not require a T1 to the Core, because they will not all use this bandwidth at the same time). Most Access Lists, compression, and encryption should be implemented here. Devices at this level are almost exclusively routers.
Core level
Concentrates all traffic that needs to transverse the network. The focus here is switching traffic as fast as possible. ATM, Gigabit, SONET, and other high-speed technologies are usually implemented here (although it is becoming more common to see higher speeds closer to the Access level, as equipment prices drop).
Subscribe to:
Posts (Atom)