In the 1990s MPLS or Multi-Protocol Label Switching emerged as an exciting new method of IP routing. At a time when traditional routing methods remained inefficient, MPLS offered users a more efficient way to send packets to IP addresses. Unlike a service that can be installed, MPLS is most accurately described as a technique. This technique is most commonly used by entities looking to provide VPNs and traffic engineering.
MPLS has become popular as a technology used to enhance Ethernet connectivity. As scalability and reliability become bigger concerns for enterprises, MPLS has offered users a way to prioritize connections within a service. Today any organization looking to increase their network efficiency and scalability would be well advised to consider using MPLS.
Over the past few years, there has been an attitude that MPLS is becoming outdated and will be replaced by more efficient technologies like SD-WAN. However, we argue that this is far from the case. In this article, we look at what MPLS is and why it is here to stay for the long term.
A Brief Overview of MPLS
In most networks each router decides on the route passing packets will take. At each, the routers provide an IP lookup to find where next to send the data. MPLS uses label switching and finds the end router to set a route straight to the end location. Routers then read this label to pass the packets straight to their destination. As a result, routers throughout the network don’t need to conduct IP lookups because all the information is already there.
An In-Depth Look at MPLS
On a traditional IT network, whenever a router receives an IP packet it is provided with a destination IP address. This tells the router where the packet’s final destination is. While this seems pretty reasonable on the surface, it isn’t conducive to efficiency. The reason is that the router has no information about how the packet should travel to its destination. In other words, traditional IP routing provides a limited amount of information on the route a packet should take.
The MPLS solution to this problem is to make the first router that intercepts a packet to be the one that decides its future route. The first route to make contact gives each packet a label which can be read by routers further down the chain. Crucially, packets are forwarded at the switching level rather than the router level. This results in lower transfer speeds and less hardware usage.
MPLS sits between the second and third layers of the OSI model. Layer 2 is used for protocols like Ethernet which are used to transport packets and Layer 3 covers the actual routing of packet data. MPLS is used to interlink the two and serves to speed up the transfer process.
At its most basic, an MPLS network is connected to a cloud service that connects to every node within your network. Essentially, MPLS acts as a VPN. MPLS is either a point-to-point VPN, Layer 2 MPLS VPN, or Layer 3 MPLS VPN. Whereas point-to-point connectivity needs routers on both sides of the network to work, MPLS doesn’t need any additional hardware.
MPLS acts almost like a bookmark. When a router uses MPLS its routing table is broken down with each section given a unique number. In technical terms, the Label Edge Router (LER) provides every packet with a label which is used to identify a Forwarding Equivalence Class (FEC). LERs also have the responsibility of removing this label at the exit point of the network and replacing it with a normal IP address.
Whenever the LER receives a packet without a label, the LER needs to assign it with an MPLS label. Once the packet has been labeled it is then sent to the next Label Switch Router (LSR) in the chain. Once the LSR receives the packet, it scans the MPLS label in the header and does one of two things; it changes the MPLS label and passes it on or if the packet is ready to leave the MPLS network then the LSR removes the MPLS label altogether. Once the latter is done the next node reads the routing information to send it to its final destination.
Once a label is assigned to a packet, it is sent to its next destination down the Label-Switched Path (LSP). The LSP is a predefined path through which your packets travel. Every router in the network needs to have a clear perspective of the LSP in order to forward packets to their next destination efficiently. When an LSR intercepts a packet, it inspects the label before sending it down the LSP to its next destination.
The main advantage to MPLS is that once a connection is made, the connecting router doesn’t have to crawl through the packet’s information before sending it to the next device, it can simply use the header instead. It provides routers with all the information they need to identify where a packet needs to be forwarded or routed to. The end result is faster packet transfers.
Devices throughout the network read the MPLS label of transferred packets in order to identify the end location to which it is being sent. In contrast, IP would send data packets but allow individual packets to decide their own path. Rather than traveling a physical path like IP traffic, MPLS uses virtual paths to get packets to their final destination.
MPLS Router Roles/Positions
The Label Switch/Router (LSR) is a router(s) that routes packet transfers using the MPLS label. This is the router that labels packets for the rest of their journey. Generally, LSRs are located in the middle of the MPLS network. Once a packet is received it determines the next location on the label-switched path and adds a label to correlate with that. It removes the old label and replaces it with a new one.
Label Edge Router
A Label Edge Router (LER) is a router situated on the end of a MPLS network which acts as an entry or exit point. LERs place labels on incoming packets before sending them to the MPLS domain. If a packet is going outbound towards the exit, then the LER removes the label and forwards the packet using the IP protocol.
In a VPN environment that is operating over MPLS, routers that are functioning as entry and exit points for the VPN are referred to as Provider Edge Routers (PER). Those routers with the sole responsibility of transferring packets are called provider routers.
Label Distribution Protocol
The Label Distribution Protocol (LDP) is used to distribute labels between LERs and LSRs. LSRs interact with each other regularly in order to exchange labels and routing information with each other to help develop their understanding of the network and making packet transfers easier.
The Customer Edge (CE) is the device on the customer’s end that a router or PE router talks to. The CE takes communications from the customers’ side and transports them straight to that of the provider. The CE router also connects to the customers’ network. The CE is at the epicenter of exchanging packets with your customers.
What is MPLS VPN and how is it used?
In many cases, you’ll hear MPLS referred to within the context of a VPN. The reason is that MPLS has the capacity to support VPN services. MPLS VPNs come in the form of Point-to-Point, Layer 2 MPLS VPN (also referred to as Virtual Private LAN Service or VPLS) and Layer 3 MPLS VPN.
Point-to-Point – This is a point-to-point connection operating at layer 2 of the OSI model through the use of LDP. This service uses virtual leased lines (VLL) to connect two different sites together.
MPLS Layer 2 VPN (VPLS) – VPLS is a layer 2 VPN which connects one point to a multipoint through the use of Ethernet. Organizations use VPLS to connect geographically separate LAN networks together. This layer uses a LDP based signaling technique from Cisco. Both Frame Relay and Ethernet can be transported through MPLS across Layer 2.
MPLS Layer 3 VPN – This is the type of MPLS service that most people refer to when they refer to MPLS VPN. In this service, administrators create virtual routing and forwarding technology on their PER. Virtual routing and forwarding mean that multiple segments of a routing table can run within one router at a time.
MPLS VPN and Cloud Services
One of the most popular applications of MPLS VPN is that of cloud services. Combining cloud services with a MPLS CPN creates a virtual private cloud. This private cloud is secure and separate from the public internet. One of the main reasons organizations have been adopting MPLS VPN for cloud services is because they can control traffic priority.
As such, MPLS VPN driven cloud services are more reliable. For example, if one application or connection is taking up too many resources it can simply be deprioritized to make way for more important processes. This provides enterprises with a much higher standard of scrutiny and discrimination than is available on the public internet. It also has the advantage of enabling an enterprise to upscale quickly. MPLS VPN can be upscaled much more easily than a traditional carrier service.
Why do I need to use MPLS?
Many organizations opt to use MPLS because of its scalability. MPLS doesn’t need any additional physical hardware in order to work which means that when you upscale you don’t need to purchase any expensive equipment. For larger organizations, this can save a lot of money over the long-term and minimize the complications that come with configuring new equipment each time the network increases in size.
Another reason why companies choose to deploy MPLS is because of its flexibility. The ability to reroute traffic according to the most efficient route and minimize disruptions is very useful. Traditional IP routing may let packets choose their own destination, but this doesn’t offer the speed that an MPLS fast-track packet transfer does. MPLS is also flexible in the sense that your service provider can provider layer 2 and 3 VPNs in one place.
Finally, you have increased performance because of label-switching. Changing the route of packet transfers at the switching layer means that devices down the chain can pass packets on more efficiently. As mentioned above this results in lower speeds and less hardware usage. This is particularly advantageous in larger organizations that are conducting lots of different packet transfers.
An MPLS chooses the route your traffic takes, which means it can avoid congested routes in favor of optimal paths. This is a big advantage because it means that your transfers don’t have to collide with each other and affect your organizational performance.
Flexible routing makes the process of rerouting traffic incredibly fast. This makes things easier for individual packets and increasing the performance of the network as a whole. Voice-based services and video applications are two areas where the quality of service is incredibly important to prevent unnecessary delay.
What are the Disadvantages of MPLS?
Even though you don’t have to worry about configuring your hardware, you take on a new concern in managing your relationship with your ISP. Your network provider is responsible for providing you with an MPLS cloud and as such you will have to work with the provider to make sure that your MPLS traffic is routed correctly. This means that you have to hand over partial control of your network. This is a considerable drawback as many organizations will be dealing with information they want to keep private.
This is also problematic because it means MPLS isn’t completely secure. An MPLS doesn’t have any features in place to protect your data. This means that once it’s up and running you are open to more external threats. This can be mitigated by ensuring that your devices are properly secured but it is something to think about before pulling the trigger on an MPLS environment. One of the most common methods organizations use to get around this issue is by encrypting all traffic transferred between two routers.
MPLS vs SD-WAN
While MPLS is still widely-used it is anticipated by many that SD-WAN (Software-Defined Wide Area Network) is going to take over in the future. An SD-WAN is applied to standard WAN connections in order to connect devices over a long distance. Generally, these are used by large corporations or data center providers. It is most well-known for helping to support cloud services like Salesforce and Office 365.
One of the biggest advantages SD-WAN has over MPLS is higher performance. SD-WAN uses a combination of MPLS, broadband, and LTE to stay connected. In effect, this creates a hybrid network that can be switched between depending on the speed of packet transfers and real-time network performance. In practice, this results in better packet delivery.
That being said MPLS is not far behind in terms of its reliability. It is an efficient method of delivering packets and provides a high quality of service. The problem is that MPLS is operated on a shared network which often results in competition over bandwidth. This can be a considerable cause of congestion when compared to SD-WAN.
With regards to security, MPLS does offer some protection but handling by ISPs runs the risk of data being shared with third parties. This is exacerbated because MPLS isn’t encrypted either. In contrast, SD-WAN acts more like a VPN and allows you to send information without it being passed on to third parties. This means that SD-WAN has the edge in terms of security.
Even though SD-WAN has a leg up on MPLS, it is really only needed if you’re running cloud services. However, if you’re simply looking to connect without the use of cloud services, then MPLS has more than enough baseline capabilities to be worth your time. Of course, if you’re not comfortable with your data being handled by your ISP then SD-WAN might be the better choice.
See also: WAN optimization
MPLS is Here to Stay
If you’re serious about making your packet routes more efficient and increasing your network’s performance, then MPLS is something you should definitely consider. Larger organizations who are constantly having to upscale their technical infrastructure will benefit from an MPLS because it will reduce the need to purchase new hardware. This will help to reduce overhead costs dramatically.
While this comes at the cost of some of your privacy and having to work with your network provider, the benefits are more than worth the sacrifices. MPLS has its supporters and its detractors, but its benefits are clear to see. It has the ability to support scalability and reliability of service in a way that traditional IP routing connections cannot.
The rise in the use of Ethernet and Wide Area Network services suggests that MPLS is as popular as ever. No matter what the cynics may say, most users gravitate towards Ethernet technology more than any other alternative. As long as Ethernet remains the main choice of connection, MPLS will be in the background.