Why is OSPF outdated in the competition with IS-IS

Latest reply: Feb 20, 2020 12:21:23 119 4 6 1

Hi friends,

Last year, I wrote an article comparing OSPF and IS-IS. This article will explain why OSPF is outdated from the perspective of protocol development history.

OSPF is a protocol invented when the IETF and OSI are in competition. OSPFv1 (1131) is not a stable protocol and is not widely deployed. OSPFv2 (2328) is a relatively stable and widely deployed protocol and defines a set of LS types, as follows: 

isis

However, OSPFv2 has a description of the processing of unknown LS types, which is considered as a failure and lesson in protocol design.

Unknown LS Type messages use the flooding mechanism, similar to the BGP attributes of transtive and non-transtive, which is a separation of "message content" and "message delivery mechanism". It's an opaque mechanism and an idea. In the protocol design, BGP / LDP has a similar mechanism.

Examine the LSA' s LS type. If the LS type is unknown, discard the LSA and get the next one from the Link State Update Packet.This specification defines LS types 1-5 (see Section 4.3). 

With the emergence of RSVP-TE, especially SR (2013), IGP needs to carry some additional information. OSPFv2 extends the opaque LS types, which is defined in RFC7684 in TLV mode. 

isis

 OSPFv2 is not applicable to IPv6 because OSPFv2 is a specific and independent protocol of IPv4, which is another OSPFv2 error. Therefore, OSPFv3 is introduced.

The first version of OSPFv3, 2740, comes with the first wave of IPv6 (IPv6 headers were defined and stabilized in 199x). It defines another set of LS types, but still uses fixed formats.

 

isis


Although the S2740 is compatible with OSPFv2, and the processing of Unknown LS Type is also modified, OSPFv2 and OSPFv3 are generally considered as two different protocols.

2.9. Handling unknown LSA types

Handling of unknown LSA types has been made more flexible so that, based on LS type, unknown LSA types are either treated as having link-local flooding scope, or are stored and flooded as if they were understood (desirable for things like the proposed External-Attributes-LSA in [Ref10]). This behavior is explicitly coded in the LSA Handling bit of the link state header's LS type field (see Section A.4.2.1).

The IPv4 OSPF behavior of simply discarding unknown types is unsupported due to the desire to mix router capabilities on a single link. Discarding unknown types causes problems when the Designated. Router supports fewer options than the other routers on the link.

The second version of OSPFv3, 5340, is introduced with the second wave of IPv6 (commercial deployment of IPv6 in 2008). It enhances the functions of RFC 2740 (as listed in Chapter 3) and retains the LS Type type and format defined in RFC 2740. This version is implemented by various vendors in the early commercial deployment phase of IPv6. The OSPFv3 described in our manuals is also RFC 5340. RFC 5340 is compatible with the RFC2740.

3. Differences with RFC 2740

OSPFv3 implementations based on RFC 2740 will fully interoperate with implementations based on this specification. There are, however, some protocol additions and changes (all of which are backward compatible).

However, it is difficult to import the Opaque LSAs into OSPFv3. For reasons such as flooding scope, if the Opaque LSA of OSPFv2 is directly defined in OSPFv3, it will overlap with the existing LS Type of OSPFv3 in terms of semantics and scope.

The big mistake made by OSPFv3 is that the LS Type field is still in a fixed format, though the Opaque LSA of OSPFv2 already uses the TLV+SubTLV mode. This led to another change in OSPFv3. RFC 8362 defines a new set of LS type. Except for the fixed part, all the contents in the LS type adopt the TLV format. OSPFv3 extensions of SR/SRv6/BIER are based on RFC 8362.

isis

 The following is the comparison between RFC 8362 and RFC 5340 TLVs.  

isis

The following lists the TLVs defined in RFC 8362 .

 

isis

Although OSPFv3 RFC 8362 supports SR/SRv6/BIER, it is too late. Most venders implement SR/SRv6 and new features using IS-IS. 

So here's a new question: Will OSPF be eliminated? Time will give the answer. Let's review the history: Why did OSPF fail? Why IS-IS is so powerful? This is a tragedy! Because they have the same original design idea, OSPF is introduced after IS-IS is launched and use IS-IS for reference.

IS-IS is designed based on OSI. The lay-3 of the well-known 7-layer model includes the network address and network layer routing protocol. Its network address is variable-length (maximum 20 bytes) NSAP address. The network-entity is actually the network layer address defined by the ISO. In other words, the ISO-designed IS-IS protocol is also Protocol Specific and does not consider supporting IP / IPv6.

Another unique point of IS-IS is that it strictly complies with the layered protocol model. The IS-IS protocol only relies on L2 and does not depend on the identification system of L3.

If the IS-IS protocol is on an Ethernet link, it is a directly defined Ethernet frame format and does not rely on the identification system and mechanism of L3.

OSPF, however, depends on the Layer 3 identification system (multicast address of packets, such as the router ID format and LSA content). The network-entity identification system is different from IPv4/IPv6 addresses, but the types of messages (hello/PDU) flooded by nodes are stable, the contents in the message are the TLV from the beginning (It’s called CLV in RFC1195). 

When extended support for IPv6 is required, or when extended support for MT or extended support for Opaque information, only the TLV in the message needs to be extended. The RFC is short, which is easy to implement. 

OSPF is designed based on IPv4 and is modeled after IS-IS. It is IPv4 specific. Many customized optimizations, such as non-zero areas, come up with Default, Stub, NSSA, etc.

IS-IS is just level-2. Therefore, OSPFv2 is very popular under IPv4. It is easy to learn and use and has a good identification system and powerful functions.

However, too much IPv4 Specific and IPv4 Specific Optimization are a drag on OSPF, and switching IPv6 in the future is a burden.

In addition, more important is the detailed basic mechanism, such as the extension mechanism (using TLV, handle unknown messages) and message mechanism (using IP packets instead of Layer 2 packets). Both mechanisms fail. This is reflected in the evolution of IPv4 -> IPv6 -> Opaque LSAs over the past decades.


This is what I want to share with you today, thank you!

  • x
  • convention:

umaryaqub
MVE Created Feb 18, 2020 05:36:19 Helpful(0) Helpful(0)

Thanks for sharing
  • x
  • convention:

A%20network%20professional%20eager%20to%20learn%20and%20help.I%20have%208%20years%20of%20network%20experience%20and%20I%20am%20working%20with%20Huawei%20VAP%20and%20looking%20after%20IP%20Projects%20design.
gabo.lr
MVE Created Feb 18, 2020 13:31:38 Helpful(0) Helpful(0)

Great post!
  • x
  • convention:

Telecommunications%20and%20Electronics%20Engineer%2C%20with%208%20years%20of%20experience%20working%20with%20Huawei%20equipment.
Faridrami
Created Feb 18, 2020 16:48:11 Helpful(0) Helpful(0)

Thanks for sharing
  • x
  • convention:

ViktorG
Created Feb 20, 2020 12:21:23 Helpful(0) Helpful(0)

Hi there, @Leah21 !

Thanks for your sharing :)

Just was wondering, how many implementations have you guys met, where ISIS is working as PE to CE protocol?
What is the strong points of ISIS loop prevention in the L3VPN PE to CE environment???

If we start talking about the strong ISIS sides - would be fair to mention that OSPF is a  way better (and of course more complicated) in the world of PE to CE IP routing provisioning.

Thanks!
  • x
  • convention:

Take%20care%20and%20have%20a%20great%20day!%3Cbr%2F%3E%3Cbr%2F%3EViktor

Comment

Reply
You need to log in to reply to the post Login | Register

Notice Notice: To protect the legitimate rights and interests of you, the community, and third parties, do not release content that may bring legal risks to all parties, including but are not limited to the following:
  • Politically sensitive content
  • Content concerning pornography, gambling, and drug abuse
  • Content that may disclose or infringe upon others ' commercial secrets, intellectual properties, including trade marks, copyrights, and patents, and personal privacy
Do not share your account and password with others. All operations performed using your account will be regarded as your own actions and all consequences arising therefrom will be borne by you. For details, see " Privacy."
If the attachment button is not available, update the Adobe Flash Player to the latest version!

My Followers

Login and enjoy all the member benefits

Login and enjoy all the member benefits

Login