Skip to document

Generalized Gmpls and LSR (Explicit Routing)

Course

Systems Programming (01:198:214)

23 Documents
Students shared 23 documents in this course
Academic year: 2021/2022
Uploaded by:
Anonymous Student
This document has been uploaded by a student, just like you, who decided to remain anonymous.
Rutgers University

Comments

Please sign in or register to post comments.

Preview text

Generalized GMPLS and LSR (Explicit Routing)

Each router in an overlay network could have a virtual circuit connecting it to each of the other routers, but for the sake of simplicity, we have just displayed the circuits from R1 to all of its neighbor routers in this instance. We refer to R1 as having five routing adjacencies since it needs to communicate with each of its five routing neighbors via the routing protocol. Figure 4(b), in contrast, shows LSRs in place of the ATM switches. Virtual circuits that once connected the routers are no longer present. R1 therefore only shares one adjacency with LSR1. Running MPLS on the switches significantly reduces the number of adjacencies that each router needs maintain in big networks. It can also significantly lessen the amount of effort required by the routers to inform one another of topology changes.

The edge routers now have a complete view of the network topology as a result of using the same routing protocols on LSRs and edge routers. This means that the edge routers will have a higher probability of selecting a good new routing if a network connection or node dies than if the ATM switches rerouted the impacted VCs without informing the edge routers. It should be noted that the process of "replacing" ATM switches with LSRs is actually accomplished by altering the protocols that are running on the switches; in other words, an ATM switch may frequently be changed to an MPLS LSR by upgrading merely its software. Furthermore, in a mode known as "ships in the night," an MPLS LSR may continue to offer common ATM features while also executing MPLS control protocols.

In more recent times, optical switches and STDM devices like SONET multiplexors have been included in the idea of running IP control protocols on devices that are unable to natively forward IP packets. The term for this is generalized MPLS (GMPLS). To give routers with topological information of an optical network, similar to what was done with ATM, was one of the driving forces behind GMPLS. The lack of standardized protocols for operating optical devices was even more significant, making MPLS the obvious choice for the task. Although the capability is more frequently referred to as explicit routing than source routing, MPLS offers an easy approach to add capabilities akin to source routing to IP networks. The fact that the route is typically chosen without regard to the packet's actual source is one factor in the distinction. Usually, one of the routers inside a service provider's network is to blame. The application of MPLS's explicit routing feature is seen in Figure 4.

Because of its design, this type of network is frequently referred to as a fish network (the routers R1 and R2 form the tail; R7 is at the head). Assume that the network's operator has decided that all traffic from R1 to R7 should take the R1-R3-R6-R7 route, and all traffic from R2 to R should take the R2-R3-R4-R5-R7 route. Making efficient use of the capacity available along the two different routes from R3 to R7 would be one justification for such a decision. Consider the traffic from R1 to R7 as making up one forwarding equivalence class (FEC), and the traffic from R2 to R7 as making up a second FEC. With standard IP routing, it is challenging to send traffic in these two classes via distinct channels because R3 often doesn't consider the origin of the traffic when making forwarding decisions. If the routers are MPLS enabled, it is simple to achieve the desired routing because MPLS uses label switching to forward packets. Traffic engineering, which is the process of ensuring that a network has enough resources to meet the

demands placed on it, is one of the uses of explicit routing. An essential component of traffic engineering is managing the precise paths that the traffic takes. Through the use of a feature called quick reroute, explicit routing can also aid in strengthening networks' ability to withstand failure. It is possible, for instance, to plan a route from router A to router B that specifically avoids a particular link L. If connection L breaks, router A could direct all traffic going to B via the predetermined path. A doesn't have to wait for routing protocol packets to go across the network or for different other nodes in the network to run routing algorithms because the backup path has already been calculated and packets are explicitly routed along the path. This can, in some cases, greatly shorten the time required to redirect packets around a point of failure.

Was this document helpful?

Generalized Gmpls and LSR (Explicit Routing)

Course: Systems Programming (01:198:214)

23 Documents
Students shared 23 documents in this course

University: Rutgers University

Was this document helpful?
Generalized GMPLS and LSR (Explicit Routing)
Each router in an overlay network could have a virtual circuit connecting it to each of the other
routers, but for the sake of simplicity, we have just displayed the circuits from R1 to all of its
neighbor routers in this instance. We refer to R1 as having five routing adjacencies since it needs
to communicate with each of its five routing neighbors via the routing protocol. Figure 4.21(b),
in contrast, shows LSRs in place of the ATM switches. Virtual circuits that once connected the
routers are no longer present. R1 therefore only shares one adjacency with LSR1. Running
MPLS on the switches significantly reduces the number of adjacencies that each router needs
maintain in big networks. It can also significantly lessen the amount of effort required by the
routers to inform one another of topology changes.
The edge routers now have a complete view of the network topology as a result of using the
same routing protocols on LSRs and edge routers. This means that the edge routers will have a
higher probability of selecting a good new routing if a network connection or node dies than if
the ATM switches rerouted the impacted VCs without informing the edge routers. It should be
noted that the process of "replacing" ATM switches with LSRs is actually accomplished by
altering the protocols that are running on the switches; in other words, an ATM switch may
frequently be changed to an MPLS LSR by upgrading merely its software. Furthermore, in a
mode known as "ships in the night," an MPLS LSR may continue to offer common ATM features
while also executing MPLS control protocols.
In more recent times, optical switches and STDM devices like SONET multiplexors have been
included in the idea of running IP control protocols on devices that are unable to natively
forward IP packets. The term for this is generalized MPLS (GMPLS). To give routers with
topological information of an optical network, similar to what was done with ATM, was one of
the driving forces behind GMPLS. The lack of standardized protocols for operating optical
devices was even more significant, making MPLS the obvious choice for the task. Although the
capability is more frequently referred to as explicit routing than source routing, MPLS offers an
easy approach to add capabilities akin to source routing to IP networks. The fact that the route is
typically chosen without regard to the packet's actual source is one factor in the distinction.
Usually, one of the routers inside a service provider's network is to blame. The application of
MPLS's explicit routing feature is seen in Figure 4.22.
Because of its design, this type of network is frequently referred to as a fish network (the routers
R1 and R2 form the tail; R7 is at the head). Assume that the network's operator has decided that
all traffic from R1 to R7 should take the R1-R3-R6-R7 route, and all traffic from R2 to R7
should take the R2-R3-R4-R5-R7 route. Making efficient use of the capacity available along the
two different routes from R3 to R7 would be one justification for such a decision. Consider the
traffic from R1 to R7 as making up one forwarding equivalence class (FEC), and the traffic from
R2 to R7 as making up a second FEC. With standard IP routing, it is challenging to send traffic
in these two classes via distinct channels because R3 often doesn't consider the origin of the
traffic when making forwarding decisions. If the routers are MPLS enabled, it is simple to
achieve the desired routing because MPLS uses label switching to forward packets. Traffic
engineering, which is the process of ensuring that a network has enough resources to meet the