The Link-State Routing is a very popular routing protocol used in a WAN network and subject to number of studies. It is very important to know the performance and cost effectiveness of each routing protocol used in the various types of network such as LAN, WLAN, WAN and WWAN, etc.
This post is analyze the comparative benefits of using Link-State Routing Protocol over wired and wireless WAN network and how this can be achieved using network simulator programs such as NS2 or NetSim.
LAN vs. WAN network
A LAN is a small network that span a small area such as a building or a room. A LAN(local area network) can connect to other LAN using a wired or a wireless medium. Multiple LANs connect to each other and spread across a geographical area, it’s called a WAN or Wide Area Network.
A LAN is faster because the nodes are connected directly to each other. The distance between the WAN nodes is greater and the communication is slower. The major difference between the LAN and WAN is in layer 2 (data link) protocols and equipment.
Since, we are only concerned with WAN, here is a list of WAN protocols.
- Asynchronous Transfer Mode (ATM)
- X. 25
- Frame Relay
- Point-to-Point Protocol (PPP)
- Synchronous Optical Network (SONET)
- Synchronous Digital Hierarchy (SDH)
There are a few other protocols available.
Wireless WAN is different from Wireless Local area connection (WLAN) and employs different technology. For Example, GSM , MANET can be considered as WWAN connectivity. Sometimes the WWANs are closed network such as MANET (mobile ad hoc network).
What are routing methods ?
Routing is forwarding network packets from one location to another location in a best way.This process involves knowing the network topology and routing using a routing algorithm.The device responsible for routing is called a ROUTER.
Since the routing is very important when the network is huge such as a WAN network. There are 3 types of routing methods employed.
- Static Routing
- Dynamic Routing
Particularly , dynamic routing is very interesting. Here is a diagram that define two specific dynamic routing.
Link-State Routing (LS)
The dynamic routing protocol update the routing information on each router dynamically. Distance vector update the information periodically to it’s neighbor and adds it’s own distance and do not learn anything else. Distance Vector is not suitable for a WAN because of the limited information for routing.
The WAN is a large network and we have to consider the cost of communicating between two nodes. A Link-State Routing Protocol finds the shortest path between two nodes in a WAN network. The Link-State Routing maintains a full information about the network. Each and every node has a “MAP” of the network topology.
The router creates link state packets if there is a change in the network or nodes. LSP flooding is the method to make sure that each an every neighboring router receives the message and update it’s “MAP”. Any previously stored information on the node is overwritten if new info is received.
Performance Evaluation Method
There are different types of wired and wireless networks available today. Local Area Networks and Wide Area Network are of various size, complexity and diversity. This makes it very important to measure the performance of the network to understand its current behavior for future improvements and find solution to known issues.
Wide Area Network and Wireless Wide Area Network use the routing protocol for almost all type of communication extensively. It is very important to understand the behavior of commonly used routing protocol such as Link State Routing Protocol. This will help us in deciding which of the protocol is suitable for a particular type of Wide Area Network.
To understand the behavior of networks it important to gather data for analysis using different techniques. The data is gathered online or offline from a network and it is very important to know what we are trying to measure before we start gathering data.
This analysis leads to the development of models to study the network performance. There are three types of models used:
- Statistical Models
- Queuing Model
- Simulation Model
We cannot say that one method is superior to another because each method has its own advantages and disadvantages. Statistical Method is very good in measuring the quantitative data from the network. Queuing Model is very good for network performance analysis because it is used for measuring the response time of the network devices or other resources.
The simulation model is a cost effective way to simulate the network since the cost of creating desired network for analysis is very high. An abstract model of the network and system is generated for analysis using computer programs. It gives lot of flexibility to researcher to try different scenarios.
What to measure ?
The routing protocol has a specific kind of characteristics then the regular network activities. Any routing protocol can be evaluated using four metrics given below
- Delivery Ratio – It is number of packets received by the receiver divided by number of packets
sent from the source.
- Latency – It is the time between creating of packet to successful delivery at the destination.
- Path Length Optimality – Difference between a path in the network to send packet and the
length of the best path in the network.
- Control Overhead– Number/Size of routing control packets sent by the routing protocol which
indicates the efficiency of the routing protocol.
The routing protocols will be analyzed for following parameters
- Traffic Sent/Received
- Network Convergence
- Point-to-point delay
- IP Traffic Delay
- Process Delay
How to measure the performance using simulation ?
NetSim is the tool that you can use to measure the performance which is single process discrete event simulator. The nodes in the NetSim are finite unlike analytic models that consider the network to be infinite. In NetSim, we can create simple networks and run discrete event simulation (DES) and analyze the results.
Since the project is running on a simulated environment using NetSim and both Wired and Wireless network is covered for the project. It is possible that some of the parameter may not be available for performance analysis because of non-availability of the feature in Network Simulator tool.
Also, the performance analysis is to be conducted on a Windows OS which means any affect due to the operating system may be difficult to recognize unless specified by NetSim.
Benefits of NetSim over other performance analysis tools
The simulation software is used for analyzing the network performance over simulated networks. These simulation software are categorized into following types
- Discrete Event Simulators
- Agent-Based Simulators
- Continuous Simulators
- Hybrid Simulators
Discrete Event Simulators models the transactions flows and these models consist of entities of different types and resources that service these entities. These simulators are not suitable to read continuous movement of traffics. There are control elements to know the state of entities and resources.
Agent-Based Simulators are sub class of Discrete Event Simulators and consists of mobile agents that are different from the discrete event simulators. The agents have properties and methods which is not available with discrete event simulators.
Continuous Simulators describe the events that use differential equations and information that move continuously. This is quite opposite to discrete simulators where information flow infrequently. Hybrid is combination of both discrete and continuous simulation.
- Performance analysis of AODV, DSR, OLSR and DSDV Routing Protocols using NS2 Simulator
- Performance Analysis of Wired and Wireless Network using NS2 Simulator
- Performance analysis of Optimized Link-State Routing Protocol for Ad Hoc Network
- A Comparative Study and Simulation of AODV MANET Routing Protocol in NS2 & NS3
- OSPF for MANNET