MPLS Means Speedy Network Traffic and Easy Management for Multiple Protocols

 
MPLS speeds network traffic, supports many service models, and is easier to manage.

What is MPLS?

Multi-Protocol Label Switching (MPLS) is a wide-area network (WAN) service for high-performance telecommunication systems, and is used for Private IP VPNs. It directs and increases the speed of network data between nodes by using short-path labels to specify a direct path instead of relying on traditional long addresses. MPLS saves resources by avoiding time-consuming look-ups in the system’s routing tables.

As its name implies, MPLS works with and can bundle together packets from different network protocols such as IP, T1/E1, Frame Relay, ATM and DSL, which MPLS labels for virtual paths by using nodes instead of end-points. Multi-Protocol Label Switching makes a network run faster, and it also gives better Quality of Service (QoS) and management control over the network.

Summary benefits of MPLS

• Supports multiple protocols

• Private network without encryption bottlenecks

• Higher-priority traffic runs faster and cleaner

• Gives both better Class of Service (CoS) and better Quality of Service (QoS)

• More uptime

• Better redundancy routing because each site sees all other sites

• Robust recovery capability

• Speeds above 1M

• Automatic Protection IP-VPN Fail-Over Routing

Which companies can benefit from MPLS?

MPLS helps companies with 3+ sites that need to communicate in real time and/or send large amounts of data between locations. It’s ideal for companies with multiple sites seeking to consolidate all voice, Internet and WAN connectivity together onto the same access circuit.

How does MPLS work?

Multi-Protocol Label Switching (MPLS) labels packets and specifies a path for them, which avoids any need for routers to look up addresses for the next node on that path. It works with a wide variety of protocols including Asynchronous Transport Mode (ATM), IP, Ti/E1, DSL, Frame Relay, Ethernet and Synchronous Optical Network (SONET).

MPLS makes decisions about packet-forwarding strictly based on the label, instead of looking at the contents of the packet itself, thus saving time by identifying paths to nodes instead of end-points. This strategy allows networks to avoid relying on any single OSI data-link layer technology. When using Multi-Protocol Label Switching, there is no need to distinguish between multiple layer-2 networks. Since MPLS works at a level between a data-link layer (layer 2) and network layer (layer 3) it is sometimes called a layer 2.5 protocol.

The roots of MPLS

MPLS was created as a consolidated data service for both datagram-type packet-switching networks and circuit-based networks. MPLS offers lower-overhead service than ATM, while providing for frames of varying length. Specifically, it solves the signaling-protocol and cell-switching issues of ATM services. MPLS capitalizes on the fact that, because today’s optical networks are so fast, the small cells provided by ATM are unnecessary. As well, MPLS provides traffic management and out-of-band control features for large networks, which are similar to the benefits offered by Frame Relay and ATM.

Label-switching

The Multi-Protocol Label Switching strategy operates by labeling packets with a header consisting of at least one label, described as a label stack. Each stack has 4 fields: A twenty-bit label value, a three-bit traffic class for priority, a one-bit flag indicating bottom-of-stack, and an eight-bit time-to-live (TTL) value. It saves time and resources because packets can be switched after undergoing only a label look-up and switch, without any need for IP table look-up. Label-switch routers (LSRs) are employed to route packets based on their labels. Label-edge routers (LERs) are used to label the incoming packets and likewise pop the labels off the outbound packets. When using a Penultimate Hop-Popping (PHP) scenario these tasks may be performed by the label-switch router connected to the label-edge router.

Label distribution between the LSR and LER is accomplished by a Label Distribution Protocol (LDP). In an MPLS network, the label-switch routers communicate information about labels and also regarding the reachability of routers, so packets can be easily forwarded. This characteristic helps with redundancy and reliability, since each site can see all other sites.

Label-switched pathways (LSPs) can be created by the network owner for various purposes, including routing network traffic through specific paths and establishing IP Virtual Private Networks. In some ways, an LSP is similar to a Permanent Virtual Circuit (PVC) as used by Frame Relay or ATM networks, yet an LSP does not rely on any single layer-2 protocol. When MPLS is used for a virtual private network, an LER that serves an inflow/outflow router for a virtual private network is often called a Provider Edge (PE) router, while hardware that serves only for packet-forwarding is called a Provider (P) router. Since the work of P routers is less difficult than that of PE routers, the P equipment may be simpler and possibly more reliable.

Labeling operations

When a fresh, unlabeled packet arrives at the ingress router and must be sent into the MPLS tunnel, that router first decides which Forwarding-Equivalence Class (FEC) will be applied to that packet, and then the router attaches at least one label to the packet’s new MPLS header. Next, the packet is sent to the appropriate hop router in the tunnel. When an MPLS router receives a labeled packet, the top label is checked. According to the information contained in the label, the router can perform a pop (dispose), push (impose) or a swap task on the label stack. The process of reviewing the top label can be performed with great speed because routers are configured with look-up tables which have been prebuilt. So, it saves time and resources.

When a label is swapped, it is replaced by a new label, and then the packet is forwarded onto the path mandated by its new label. Another type of operation, used by VPNs, involves pushing a new label on top of the packet’s current label, which has the effect of encapsulating the packet within a new MPLS label; this permits hierarchical routing of packets. In the reverse operation, which is a pop, the label is stripped off a packet, thus revealing the label within; this task is described as “decapsulation.” When the popped label was the final identifier contained in the label stack, it means that packet is ready to leave the tunnel. Usually, this task is performed by the LER, although in a PHP scenario it can be done by the LSR.

As a time-saving feature, the packet’s contents under the MPLS label stack do not need to be examined during label-switching operations. This process is very fast, since the LSRs look only at the stack’s top label, and they forward a packet based only on its label information without regard for the protocol of the packet’s contents itself. Because there is no need to look at a routing table regarding protocol, the MPLS service avoids the costly IP longest-prefix match-up with each hop.

After reaching the egress router where the final label is popped, the data payload alone is left to be examined. This payload can be in the form of an IP packet or any of several other kinds of packets. Because there is no label, the egress router must already have information to help with routing the payload, in order to do so without the aid of look-up tables. Once the payload arrives, MPLS can use the existing infrastructure of Frame Relay or ATM networks, since the labeled packet traffic may be mapped by virtual-circuit identifiers.

Managing MPLS paths

There are two standardized protocols for removing or installing paths. These protocols are Label Distribution Protocol, or LDP, and RSVP-TE, which is an extension of the RSVP protocol used in traffic engineering. Also, extensions based on the Border Gateway Protocol (BGP) may be employed to manage MPLS paths.

It is important to note that because the type of data inside a packet is not identified by the MPLS header, a network operator who wishes to send data with various protocols among the same routers must set a different MPLS path so that the core routers treat those traffic types differently.

Comparing MPLS to IP, Frame Relay and ATM

One of the drawbacks of a strictly-IP network is that, even when the shortest path is congested, the service still specifies that it must be used. In contrast, an MPLS network can specify that the shortest path with available bandwidth will be used. Also, in case of a network failure at the IP layer, the time required for restoration may be intolerable for applications such as VoIP. In contrast, MPLS offers excellent local protection, including a recovery time similar to that offered by a SONET ring of 50 ms. It also offers advantages over Frame Relay, especially for clients frustrated by bandwidth overbooking by telecom providers due to under-provisioning of data services.

In comparing MPLS to ATM, it can be said that although both technologies signal connections between end-points, and each uses a form of encapsulation to carry data, MPLS offers important advantages: It can handle packets of variable lengths but ATM carries only cells of fixed length. Also, when using an ATM network the packets must be segmented and then later reassembled through an adaptation layer, which greatly increases the complexity, time and cost of the process. In contrast, MPLS simply labels the packets and quickly sends them over the network. And, although ATM and MPLS both support tunneling by allowing connections within other connections, MPLS simply stacks labels while ATM requires virtual paths. Since the virtual path indicator (VPI) and circuit indicator (VCI) must be carried in the same cell header, ATM is limited to tunneling in only a single level.

Finally, MPLS’ most important advantage over ATM is that MPLS was designed to work closely with IP, and is supported by routers with a common interface, yet the incompatibilities of ATM demand extensive adaptation for use with IP. Since networks today are mostly IP-based, and because of the benefits listed above, many network operators are migrating toward MPLS because of its overall advantages, including the ease of management control over QoS. MPLS is used to connect as few as three sites, and it is scalable up to deployments containing many thousands of sites. MPLS is especially popular for large-scale deployments in retail businesses for transaction data, and for telecommunications uses.

To learn more about the features and benefits of Multi-Protocol Label Switching, please contact AeroCom today!