EIGRP and OSPF Comparison
For
<Client>
Client Sponsor
Prepared By
Scott Hogg
Project Number
02
Date
March 14, 2002
EIGRP and OSPF Comparison
Distribution ListName / Title/Duties / Company
John Vogt-Nilsen / Manager – Network Operations / <Client>
Sammy Hutton / Principal Systems Analyst / <Client>
Scott Hogg / Principal Consultant / Lucent
Phil Colon / Managing Consultant / Lucent
Revision History
Version / Date / Author / Comments
1.0 / 03/14/2002 / Scott Hogg / Initial Draft
TABLE OF CONTENTS
Paragraph & Description Page
1.0 Executive Summary 1
2.0 Introduction 1
2.1 Industry Standard protocols vs. Proprietary protocols: 2
3.0 Technical Background 2
3.1 Types Of Routing Protocols 2
3.1.1 Static Routing 2
3.1.2 Distance Vector Protocols 3
3.1.3 Link-State Protocols 4
3.1.4 Advanced Distance Vector Protocol 5
3.1.5 Path Vector Protocols 5
4.0 Protocol Decision Criteria 5
5.0 OSPF 8
6.0 EIGRP 18
7.0 Analysis 21
8.0 Recommendation 22
9.0 References 23
9.1 URLs 23
9.1.1 OSPF 23
9.1.2 EIGRP 24
9.2 Books 24
9.2.1 OSPF 24
9.2.2 EIGRP 24
EIGRP and OSPF Comparison
1.0 Executive Summary
The <Client> network is based on the TCP/IP protocol, which permits the efficient routing of data packets based on their IP address. Cisco routers are used at various points in the network to control and forward the data. Alcatel OmniSwitch switch/routers are also used in the Site 2 facilities.
At the current point a decision is being made by <Client> on whether to keep the existing Alcatel infrastructure in the Site 2 facility or migrate that equipment to similar Cisco equipment as exists in Site 1. The current Alcatel equipment is experiencing severe problems such as hardware failures, power supply failures, operating system memory leaks resulting in reboots. If the decision is made to upgrade the Alcatel switch/routers then an evaluation will need to be made on what the proper routing protocol <Client> should be running corporate wide will be needed. This would be an evaluation of suitability of Enhanced Interior Gateway Routing Protocol (EIGRP) or Open Shortest Path First (OSPF).
In order for the routers to effectively and efficiently distribute data to the users in the field, the routers must be programmed with the topology of the network. In other words, the routers must contain a “map” of the other routers in the network and what TCP/IP devices are connected to them.
There are a number of methods to program the routers with this information and to change the program as the network changes. The choice of method, or routing protocol is a critical factor in the success of the network over time. Factors that differentiate one routing protocol from another include the speed that it adapts to topology changes (convergence), the ability to choose the best route among multiple routes (route calculation), and the amount of network traffic that the routing protocol creates.
Based on this evaluation of the suitability of a routing protocol for <Client>’s routed TCP/IP network the EIGRP routing protocol should be used in the Alcatel routers are upgraded to Cisco routers. However, if the Alcatel routers are retained for service within the Site 2 campus then <Client> has no alternative but to run OSPF throughout the organization.
2.0 Introduction
Cisco has dominated the router industry for many reasons. One of the most common reasons is Cisco’s support for a multitude of protocols as well as features in their IOS to enhance a router’s ability to control traffic and improve performance, and in some cases, save money. It makes sense for a company that utilizes Cisco routers in their network, to take advantage of the features and functionality that has helped Cisco become the leader in terms of market share.
2.1 Industry Standard protocols vs. Proprietary protocols:
A case can be made for both standard as well as proprietary protocols.
STANDARDS PRO: A standards based protocol will theoretically allow routers of different manufacturers to inter-operate.
STANDARDS CON: Standards based protocols require industry approval for changes. Historically changes, as well as improvements or advancements, are rare. Changes to the OSPF RFC have not occurred since 1986.
PROPRIETARY PRO: Owner can advance the protocol to new levels without the agreement of a consortium of companies resulting in a protocol with the latest in technological advancements.
PROPRIETARY CON: Protocol is not supported by other vendors requiring the implementation of a second protocol. Use of a proprietary protocol is only an issue internally when using multiple vendors for routers, requiring a gateway router to re-distribute routes. This is generally not an issue with external networks, since exterior gate protocols like BGP are used when connecting outside.
Proprietary protocol standards compliance is an issue because the <Client> network is currently comprised of both Cisco and Alcatel routers. The Alcatel routers support RIP v1 and v2, OSPF, and BGP-4 only. They don’t support Cisco’s proprietary EIGRP.
In making a determination as to which routing protocol (stay with RIP v1/v2, OSPF, or EIGRP) should be used, <Client> has to look at technical as well as the administrative benefits to be derived from each. It is obvious that RIP v1 needs to be eliminated and that decision has already been made, so a matrix of features and benefits between OSPF and EIGRP needs to be developed.
3.0 Technical Background
3.1 Types Of Routing Protocols
3.1.1 Static Routing
The simplest form of routing is static routes. The routing information is preprogrammed by the network administrator. When changes to the network occur, the route information must be manually changed throughout the network.
There are a number of advantages to using static routes. Static routing is very resource efficient, as it routing uses no additional network bandwidth, doesn't use any router CPU cycles trying to calculate routes, and requires far less memory. It is also the most secure form of routing protocol.
However, there are a number of disadvantages to static routing that eliminate it as a viable alternative on the <Client> network. First and foremost, in the rapidly changing topology of a wireless network, it is impractical for a network administrator to manually program the routing changes as they occur. Secondly, in the case of a network failure, static routing is usually not capable of choosing alternate paths.
3.1.2 Distance Vector Protocols
Distance vector protocols such as Routing Information Protocol (RIP), Interior Gateway Routing Protocol (IGRP), Internetwork Packet Exchange (IPX) RIP, IPX Service Advertisement Protocol (SAP), and Routing Table Maintenance Protocol (RTMP), broadcast their complete routing table periodically, regardless of whether the routing table has changed. This periodic advertisement varies from every 10 seconds for RTMP to every 90 seconds for IGRP. When the network is stable, distance vector protocols behave well but waste of bandwidth because of the periodic sending of routing table updates, even when no change has occurred. When a failure occurs in the network, distance vector protocols do not add excessive load to the network, but they take a long time to reconverge to an alternate path or to flush a bad path from the network.
Distance Vector Routing protocols are dynamic. Routers that use distance vector routing share information, or a routing map, with other routers on the network. As changes to the network occur, the router with the change propagates the new routing information across the entire network.
In routing based on distance-vector algorithms, routers periodically pass copies of their entire routing table to routers that are their immediate neighbors. Each recipient of this information adds a distance vector (it’s own distance value) to the routing table before it forwards it on to its neighbors. This process continues in an omni directional manner among connected routers. Eventually each router on the network learns about all the others and is able to develop a cumulative network “map.” Each router then knows how to reach any other router, and any other network connected to the router.
Distance vector routing provides a tremendous advantage over static routing. Routers are able to discover the state of the network, and to propagate changes as they occur. The most common, and most ubiquitous of distance vector routing protocols is the Routing Information Protocol, or RIP.
However, there are also some disadvantages to distance vector routing that preclude its use on the <Client> network:
· Because distance vector routing protocols periodically transmit the entire routing table to all immediate neighbors, they can add significant traffic. This is particularly problematic on a wireless network with limited bandwidth.
· Distance vector protocols are notoriously slow to converge, or adapt to network topology changes. After a change to the network, and before all the routers have converged, there is the probability of routing errors and lost data.
· Distance vector routing protocols base their routing decisions on distance, or the number of “hops” from one network to another. It does not take into consideration the speed or bandwidth of a network path. Therefore, routers may route traffic through paths that are suboptimum.
3.1.3 Link-State Protocols
Link-state routing protocols, such as Open Shortest Path First (OSPF), Intermediate System-to-Intermediate System (IS-IS), and NetWare Link Services Protocol (NLSP), were designed to address the limitations of distance vector routing protocols (slow convergence and unnecessary bandwidth usage). Link-state protocols are more complex than distance vector protocols, and running them adds to the router's overhead. The additional overhead (in the form of memory utilization and bandwidth consumption when link-state protocols first start up) constrains the number of neighbors that a router can support and the number of neighbors that can be in an area. When the network is stable, link-state protocols minimize bandwidth usage by sending updates only when a change occurs. A hello mechanism ascertains reachability of neighbors. When a failure occurs in the network, link-state protocols flood Link-State Advertisements (LSAs) throughout an area. LSAs cause every router within the failed area to recalculate routes. The fact that LSAs need to be flooded throughout the area in failure mode and the fact that all routers recalculate routing tables constrain the number of neighbors that can be in an area.
Link state routing protocols, like distance vector protocols, are dynamic. They propagate route information across networks. However, they have a number of advantages over distance vector protocols.
One of the major advantages of link-state routing is that they calculate the best route for data based on cost rather than distance. The algorithms used to determine cost vary from protocol to protocol, but it is generally based on a link’s bandwidth. Thus, the router that the data packet takes to get to its destination is optimized.
Additionally, link state protocols do not transmit their entire topology database across the network on a periodic basis. Once the network has converged, protocol traffic is limited to changes in specific links (link state advertisement packets) and keep-alive or “hello” packets.
Finally, convergence times for link state protocols are generally much shorter than for distance vector protocols. A network based on link-state routing will recognize and adapt to failures and changes much more quickly.
There are a few disadvantages to link state routing protocols that must be considered. They are generally much more complex than either static routes or distance-vector routing. This translates into higher implementation costs, higher CPU utilization, and greater memory requirements.
3.1.4 Advanced Distance Vector Protocol
Enhanced Interior Gateway Routing Protocol (EIGRP) is an advanced distance vector protocol that has some of the properties of link-state protocols. Enhanced IGRP addresses the limitations of conventional distance vector routing protocols (slow convergence and high bandwidth consumption in a steady state network). When the network is stable, Enhanced IGRP sends updates only when a change in the network occurs. Like link-state protocols, Enhanced IGRP uses a hello mechanism to determine the reachability of neighbors. When a failure occurs in the network, Enhanced IGRP looks for feasible successors by sending messages to its neighbors. The search for feasible successors can be aggressive in terms of the traffic it generates (updates, queries and replies) to achieve convergence. This behavior constrains the number of neighbors that are possible.
3.1.5 Path Vector Protocols
There is really only one Path Vector routing protocol and it is Border Gateway Protocol version 4 (BGP-4). This is the primary routing protocol used on the Internet to share routing updates between Autonomous Systems (AS). An Autonomous System is a network under a single administrative and technical control. ASs are typically defined by the boundaries of a single company or organizational entity. BGP-4 is typically used between Internet Service Providers (ISPs) and between companies and the multiple ISPs they use for upstream Internet connectivity. BGP-4 routers operate in either External BGP (EBGP) or Internal BGP (IBGP) configurations depending on whether the connectivity is between ASs or within ASs respectively. Since <Client> currently default routes toward their Internet points of presence there is little reason for <Client> to use this protocol. Regardless, BGP-4 would not be used within the corporate network and only in the future would it be used in a limited capacity at the Internet edges of the <Client> intranet.
4.0 Protocol Decision Criteria
In order to conduct a proper evaluation <Client>’s requirements for a routing protocol should be documented.
Simplicity of configuration is a significant requirement for <Client>’s selection of a routing protocol. It must be easy to configure and easy to maintain. <Client>’s IT resources are currently stretched thinly and complexity of a routing protocol is a primary consideration. Currently within the Cisco portions of the <Client> network OSPF is being used but only with all routers being in a single Area 0. This was done for simplicity and to reduce the complexity of configuring Area Border Routers (ABRs). However, the entire advantage of OSPF’s hierarchy is not being taken advantage of.
<Client> is using RFC1918 addresses internally such that Site 1 Arizona and
Western regions of the company uses the 172.16.0.0/12 while Site 2 and other parts of the company use 10.0.0.0/8. The 192.168.0.0/16 is being used for the internal side of the Internet portals. When considering a TCP/IP routing protocol the IP addressing plays a significant role in the decision and engineering process. Therefore, it is a requirement that <Client> use a routing protocol that supports Variable Length Subnet Masking (VLSM). The RIP version 1 that is being used in Site 2 in a classful routing protocol that does not support VLSM and it has already been determined that RIP needs to be phased out.