If you are part of an enterprise architecture team that supports multiple remote sites, part of your job requires understanding the different wide area network (WAN) options available to you and how they differ from each other. One highly popular technology used in today’s enterprise WAN architectures is Multiprotocol Label Switching (MPLS).
MPLS is a data transport mechanism that is designed to increase speed and reduce memory overhead in extremely large networks, such as within global WAN service providers. Instead of routing packets based on IPv4 or IPv6, as is the case with most enterprise network designs, MPLS uses less complex labels to route traffic to its intended destination. MPLS works well for WAN transport because it can work with telecommunication access technologies such as point-to-point T1/E1’s, Frame Relay, ATM and DSL. From a customer perspective, MPLS is physically handed off as either a serial interface such as a T1 -- or more likely as a standard Ethernet connection.
In this guide, we will cover how MPLS works and why it might be a better option compared to alternative WAN technologies. We’ll also touch on the difference between MPLS and its close relative, VPLS.
(Image: geralt/Pixabay with modification)
Unless you work for a WAN service provider, you dont really need to know how the MPLS cloud works at the micro-level. But to understand and appreciate the benefits inherent in the technology, its good to have an understanding of the basics. To help with that, were going to use a common enterprise MPLS deployment architecture, using the network diagram shown above. Our scenario shows an enterprise company that is leveraging an MPLS provider to interconnect three geographically dispersed locations.
The three routers colored red and labeled CE-1, 2 and 3 designate the customer edge equipment. This is likely a company-owned router or switch used to connect to the service provider network. The green routers labeled PE-1, 2 and 3 are the MPLS providers premise equipment MPLS switches. These are the first devices your company packets reach as they enter the MPLS network. Finally, the orange colored hardware in the middle of the MPLS cloud represents other MPLS Label Switch Routers (LSRs) that create the mesh of hardware, ultimately interconnecting your (and many other customer) sites together.
CE network components have no concept of what MPLS is. Instead, these devices only know how to route IPv4 traffic. And to route data from CE-1 to CE-2 or CE-3, IPv4 is exactly what CEs will use to send data across the MPLS network. Each CE router is connected to the PE router in a point-to-point connection using IPv4. IP routes to destinations are learned through static routes or learned dynamically using BGP.
When the packet eventually reaches the destination PE device, the MPLS/IPv4 knowledgeable router will pop the MPLS tag (since the CE wouldnt understand what to do with it) and forward the now normalized IP packet to the CE device.
There are a few benefits to using MPLS over other WAN options such as point-to-point links and creating site-to-site VPN tunnels across the public Internet. The first benefit is the ability to use quality of service (QoS) tags across the entire MPLS network. Time-sensitive packets (such as voice/video) can be tagged on the local network, specifying that they should receive special treatment when congestion on the network occurs. MPLS PE routers can read the QoS markings and make MPLS switching decisions based on how the packets are marked. Enterprise customers can buy different MPLS QoS tiers based on how they want their data to be handled. Oftentimes, MPLS providers offer three to four different tiers ranging from voice/video tiers which are switched over the lowest-latency and least congested paths to a best-effort tier for data that isnt latency sensitive.
Another key MPLS feature is observed when more than two remote sites are connected to each other on the same MPLS cloud. In traditional point-to-point architectures, all circuits from remote sites would terminate back at the main headquarters. To fully mesh all remote sites would likely cost too much. So when one remote site needs to send data to the other, it would have to pass it to the headquarters router first in a hub-and-spoke design. This causes slowdowns and inefficiencies. But with multiple MPLS sites, each site can send data directly in a far more efficient point-to-multipoint architecture.
Lastly, MPLS provides great scalability. Need to spin up a new site? Just add another insertion point into the MPLS cloud and youre good to go. Its just that easy.
In this guide, we showed how MPLS works at layer 3 from the customers perspective. In our example, each MPLS site would have to route its IPv4 traffic across the MPLS cloud. Virtual Private LAN Service (VPLS) is essentially MPLS, but changes it to act as a layer 2 LAN. In our example, if the three sites used VPLS as opposed to MPLS, the three sites would look as if they were connected together with a layer 2 switch. In reality, the VPLS cloud simply masks the MPLS network and creates a virtualized tunnel at layer 2.
The primary benefit here is that customer network architects can assign their own IP addressing and even run interior routing protocols across the network such as EIGRP or OSPF. Keep in mind, however, that VPLS is for Ethernet-only handoffs and is far more limited in scalability compared to MPLS. Most service providers offer both MPLS and VPLS, so its up to your network architects to determine which one -- or a combination of the two -- works best for you.
Subsea cable alternatives can provide a level of comfort for those concerned about disruptions. Realistically, they will continue to play a small, but critical role in the short term.
Maintaining existing network infrastructures often incurs huge technical debt before newer technologies are adopted over time.
Amid ongoing regional uncertainties, telecom infrastructure in the Middle East has expand to include terrestrial low-latency network infrastructure, which offers resilience and agility in navigating political complexities without relying solely on undersea networks.
