OSPF protocol
OSPF (Open Shortest Path First) is a link state-based Interior Gateway Protocol developed by the IETF organization. Currently, OSPF Version 2 (RFC2328) is used for the IPv4 protocol; OSPF Version 3 (RFC2740) is used for the IPv6 protocol.
The main features of the OSPF protocol are:
- OSPF uses multicast to send and receive messages, which can reduce the impact on other routers that do not run OSPF.
- OSPF supports typeless inter-domain routing (CIDR).
- OSPF supports load balancing of equal-cost routes.
- OSPF supports packet encryption.
- OSPF allows the entire routing domain to be divided into multiple areas to improve scalability and management. One of the areas must be the backbone area (Area 0), and the other areas must be connected to the backbone area through the backbone area or virtual connections. Communication between areas requires area border routers (ABRs).
- OSPF has five types of routers: Internal Router (IR), Area Border Router (ABR), Autonomous System Border Router (ASBR), Backup Designated Router (BDR) and Designated Router (DR).
- OSPF has five types of networks: point-to-point network, broadcast network, non-broadcast multiple access network (NBMA), point-to-multipoint network and virtual link.
- OSPF has seven types of link status advertisements (LSAs): Router-LSA, Network-LSA, Network-summary-LSA, ASBR-summary-LSA, AS-external-LSA, NSSA LSA and Opaque LSA. Different types of LSAs are propagated in different types of areas.
- OSPF has five types of messages: Hello message, Database Description message, Link State Request message, Link State Update message and Link State Acknowledgment message. These messages are used to establish and maintain neighbor relationships, exchange link status information, and acknowledge received information.
- OSPF uses Router ID to uniquely identify a router. The Router ID is a 32-bit unsigned integer that can be configured manually or automatically selected from the largest IP address of the interface.
- OSPF uses cost as a metric to calculate the shortest path to the target network. The cost can be determined based on the bandwidth of the interface or manual configuration. By default, cost=100M/interface bandwidth.
IS-IS protocol
IS-IS (Intermediate System to Intermediate System) is a link state-based interior gateway protocol developed by the ISO organization. Originally designed for CLNP (Connectionless Network Protocol), it was later expanded and modified to support both IP and CLNP networks, called integrated IS-IS or dual IS-IS.
The main features of the IS-IS protocol are:
- IS-IS works at the data link layer rather than the network layer, so it does not rely on IP addresses or CLNP addresses. IS-IS uses ISO addresses to identify network entities and services.
- IS-IS divides the entire routing domain into two layers: Layer 1 (L1) and Layer 2 (L2). The L1 area is equivalent to the normal area of OSPF, and the L2 area is equivalent to the backbone area of OSPF. L1 areas must be connected through the L2 area, and L1 areas cannot be directly connected.
- IS-IS has three types of routers: Layer 1 routers (L1), Layer 2 routers (L2), and Layer 1/Layer 2 routers (L1/L2). L1 routers can only route within the same L1 area, L2 routers can only route within the L2 area, and L1/L2 routers can route between the two levels, which is equivalent to OSPF's ABR.
- IS-IS has four types of link state messages (LSP): Level 1 LSP, Level 2 LSP, Level 1-2 LSP and external LSP. Different types of LSPs propagate in areas at different levels.
- IS-IS has four types of messages: IS-IS Hello messages, complete sequence number messages (CSNP), partial sequence number messages (PSNP), and link state messages (LSP). These messages are used to establish and maintain neighbor relationships, exchange link status information, and acknowledge received information.
- IS-IS uses System ID to uniquely identify a router. The System ID is a 48-bit unsigned integer, usually chosen from the interface's MAC address.
- IS-IS uses default metric values to calculate the shortest path to the destination network. The default metric can be determined based on the bandwidth of the interface or manually configured.
Similarities between OSPF protocol and IS-IS protocol
The OSPF protocol and the IS-IS protocol are both interior gateway protocols based on link status. Both use the shortest path first algorithm for route calculation. Both support classless inter-domain routing (CIDR). Both allow the entire routing domain to be divided into multiple Areas all use a flooding mechanism to disseminate link status information, both require the election of a designated router (DR or DIS) to reduce communication overhead in the broadcast network, and both support load balancing on equal-cost paths.
The difference between OSPF protocol and IS-IS protocol
There are also some significant differences between the OSPF protocol and the IS-IS protocol, which are mainly reflected in the following aspects:
| aspect | OSPF | IS-IS |
|---|---|---|
| work level | Network layer | data link layer |
| Address structure | Use IP address | Use ISO address |
| Router ID | Use Router ID | Use System ID |
| Regional division | Backbone area and ordinary area | First level area and second level area |
| Router type | Internal, ABR, ASBR, DR, BDR | L1、L2、L1/L2、DIS |
| Link status information | Use LSA | Use LSP |
| Message type | Hello、DD、LSR、LSU、LSAck | IIH、CSNP、PSNP、LSP |
| metric | Use cost | Use default metric values |
| Network Type | Point-to-point, broadcast, NBMA, point-to-multipoint, virtual link | Point-to-point, broadcast, NBMA |
Application scenarios of OSPF protocol and IS-IS protocol
Both the OSPF protocol and the IS-IS protocol can be applied to large networks, but they each have their own advantages and disadvantages based on different network characteristics and requirements. Generally speaking:
- The OSPF protocol is more suitable for campus networks or enterprise networks because it supports more network types, more area types, more flexible metric configuration, and more detailed routing policies and policy routing.
- The OSPF protocol also interoperates more easily with other routing protocols because it uses IP addresses for router identification and network addresses. The OSPF protocol is also easier to troubleshoot because it supports message encryption and authentication, as well as more debugging commands and tools.
- The IS-IS protocol is more suitable for carrier networks or Internet backbone networks because it works at the data link layer, does not rely on IP addresses or CLNP addresses, and can support multiple network protocols at the same time. The IS-IS protocol is also simpler and more efficient because it has only two layers of area division, only one area border router, only four link state messages, and only one metric value. The IS-IS protocol is also more stable and scalable because it uses the ISO address as the router identification and network address, which can avoid the problem of duplicate or inconsistent IP addresses and can support larger network scales.
Advantages and disadvantages of OSPF protocol and IS-IS protocol
The OSPF protocol and the IS-IS protocol each have their own advantages and disadvantages, as follows:
| aspect | OSPF | IS-IS |
|---|---|---|
| Advantage | Supports more network types, area types, metric value configuration, routing policies and policy routing; is easier to interoperate with other routing protocols; supports packet encryption and authentication; is easier to troubleshoot | Works at the data link layer and does not rely on IP addresses or CLNP addresses; can support multiple network protocols at the same time; area division is simple and efficient; link status messages are few and precise; measurement values are unified and simple; use ISO addresses to avoid IP address conflicts ;Support larger network scale |
| Disadvantages | Works at the network layer and relies on IP addresses; it can only support IP networks; area division is complex and cumbersome; link status messages are numerous and complex; metric configuration is flexible but complex; using IP addresses may lead to IP address conflicts; it is difficult to support very large network size | Supports fewer network types, area types, metric value configurations, routing policies, and policy routing; is not easy to interoperate with other routing protocols; does not support packet encryption and authentication; is not easy to troubleshoot |