How Does IP Routing Work?
Whenever an IP packet is received by a router, there is no information attached to it apart from the IP address of its destination. This means every router must make its own decisions about how to forward every packet, referring to a complex routing table every time. This process repeats at every stage of the route before the packet reaches its final destination. The result is poor performance, especially for applications which are time-sensitive such as VoIP or video-conferencing.
MPLS – An Overview
MPLS or Multi-Protocol Label Switching has been around for almost twenty years. It assigns any packet entering the network to a particular FEC (Forwarding Equivalence Class) by attaching a label in the form of a short bit sequence to it. Every network router knows how to handle each FEC type, so there’s no need for the router to perform any header analysis. Every subsequent router uses this label as a form of index into a forwarding table which supplies the packet with a brand new FEC.
This functionality allows the MPLS network to deal with packets which have specific characteristics consistently. For example, packets which carry video, voice or other real-time traffic, can be mapped easily to a low-latency route – something which is very difficult using conventional routing.
Benefits And Disadvantages Of MPLS
MPLS brings many benefits including:
- Better utilisation of bandwidth
- Improved performance
- Reduced congestion in the network
- Better experience for end-users
- Invulnerable to a denial of service attack
Of course there are also some downsides. These include:
- MPLS services must be bought from a carrier. This makes it expensive when compared to sending traffic via public internet.
- It can be hard to find a provider of MPLS services that is able to deliver coverage around the world. Most providers work in partnership with other providers to provide global coverage and this is expensive.
- MPLS was not designed to handle direct cloud access.
Does MPLS Have A Future?
With these negative points in mind, it’s not surprising that many people think that MPLS has no future. However, MPLS continues to be a key element in the WAN landscape. It is still valuable for making connections to specific point-to-point locations, for example retail facilities that have POS systems and regional offices. It is also used for any real-time applications. However a new technology has now entered the market which provides a real alternative – SD-WAN. However, the reality is that both types of technology have a key role to play in WANS today.
What Is SD-WAN?
SD-WAN applies the concepts of SDN (Software Defined Networking) to the WAN. In practical terms, it means deployed SD-WAN edge devices which apply policies and rules in order to transmit traffic on the most appropriate path.
SD-WAN is capable of routing all types of traffic, and that includes MPLS. It comes with its own unique advantage too – the enterprise WAN-traffic architect is able to sit centrally and apply policies easily across every WAN device. This contrasts with MPLS, where predetermined routes have to be provisioned painstakingly. Once each fixed circuit has been set up, it’s difficult to make changes.
However, once the MPLS network has been deployed, real-time traffic performance is guaranteed. SD-WAN is capable of routing traffic on an optimally efficient path, however once the IP packets reach the public internet, no level of performance can be guaranteed.
With this in mind, the best strategy is to offload MPLS traffic as much as possible onto the open internet, but MPLS should continue to be used for any time-sensitive application which requires guaranteed delivery.