SoFunction
Updated on 2025-04-11

The way out of IP - Introduction to MPLS

With the vigorous development of the Internet, people have also put forward higher demands for applications. Due to the lack of effective traffic and network bandwidth management methods, network congestion often occurs. The inability to provide quality of service (QoS) guarantees make many applications such as voice and video, which are incompetent for current IP technologies. The emerging multi-protocol mark switching technology (MPLS) is expected to solve this problem.

How MPLS works

According to the current definition, MPLS is a tag-based IP routing method. These tags can be used to represent hop-by-hop or explicit routing and indicate quality of service (QoS), virtual private networks, and other types of information that affect how a particular type of traffic (or traffic from a particular user) is transmitted over the network.

Currently, the routing protocol selects the shortest path between a specified source and destination, regardless of whether the path is overloaded or not. With explicit routing, service providers can select the path through which special traffic passes, allowing traffic to select a path with low latency.

The MPLS protocol implements the conversion of the third-level packet switching into the second-level exchange. MPLS can use various layer 2 protocols, and the MPLS Working Group has so far standardized the markers used on frame relay, ATM and PPP links, and IEEE 802.3 local area networks. One benefit of MPLS running on frame relay and ATM is that it brings arbitrary connectivity of IP to these connection-oriented technologies. At present, the main development direction of MPLS is in ATM. This is mainly because ATM has strong traffic management functions and can provide QoS services. The combination of ATM and MPLS technologies can give full play to its role in traffic management and QoS.

A tag is a header used to forward packets. The format of the header depends on network characteristics. In a router network, the tag is a separate 32-bit header. In the ATM, the mark is placed in the virtual circuit identifier/virtual channel identifier (VCI/VPI) cell header. In the core, only the tag is interpreted, not the packet header. A very critical point for MPLS scalability is that marking makes sense only between two devices communicating.

When an IP packet enters the core of the network, the edge router assigns it a tag. From then on, the MPLS device will view these tagged information from beginning to end and exchange these tagged packets to their destination (see schematic). Due to the reduction of routing processing, the waiting time of the network also decreases, while the scalability increases.

The quality of service type of MPLS data packets can be determined by the MPLS edge router based on various parameters of the IP packet, such as the IP source address, destination address, port number, TOS value and other parameters. For IP packets reaching the same destination, different forwarding paths can be established according to their TOS value requirements to meet their transmission quality requirements. At the same time, through the management of special routes, load balancing and congestion problems in the network can be effectively solved. If congestion occurs in the network, MPLS can establish new forwarding routes in real time to share its traffic to alleviate network congestion.

Currently, there are two protocols adopted by manufacturers. One is the Constraint-based Routing Label Distribution Protocol (CRLDP), and the other is the Resource Reservation Protocol (RSVP). The tag allocation protocol (LDP) provides communication between the edge and the core device, and combines routing protocols such as OSPF, ISIS, EIGRP (enhanced internal gateway routing protocol) or BGP to allocate tags between the edge and the core device to establish a tag switching path. Currently, the MPLS Working Group uses both methods. Although choice is often a positive thing, there are two criteria for solving the same problem, which obviously brings serious interoperability problems. Many manufacturers also find themselves shouldering the burden of implementing two sets of agreements.

Manufacturers are looking forward to the market to decide which approach will succeed, but past experience has proven that users often delay adopting new technologies because of instability.

MPLS application

MPLS will help operators provide better IP services. Like AT&T, Level 3, MCI WorldCom and UUNET have begun to deploy and market their MPLS networks. Operators also face a big problem, which is: whether to use MPLS to simplify their existing ATM network, or to establish an MPLS network without ATMs. The answer to the question depends on what type of data to be transmitted. The MPLS structure system is better for packet networks, while the ATM is the most ideal for voice and bit synchronization data.

MPLS will bring more bandwidth control, throughput guarantee and virtual private network functionality. In MPLS networks, even packet voice services such as IP voice will be improved because the waiting time is shortened and congestion control is strengthened. All in all, the result of all this is to enable users to enjoy better services from service providers. Therefore, when choosing a supplier or evaluating a contract, be aware of the situation where the other party provides MPLS.

It is obvious that MPLS is mainly aimed at service providers and operators. However, it also has many characteristics that will benefit enterprise users, whether they use public or dedicated WAN services.

Virtual private network

One application that has great potential benefits for service providers is the support for VPN services. Using MPLS for VPNs is to build a VPN by using ATMs or frame relay permanent virtual circuits (PVCs) or various forms of tunneling to interconnect customers' routers.

VPNs using MPLS can provide many advantages of PVC-based mode. Customers can choose their own addressing plans that may or may not overlap with other customers or service providers’ plans. Every customer can trust that data will only be sent to sites within that customer's VPN. Because of this, encryption is often unnecessary, unlike many tunneling methods. However, unlike PVC mode, the MPLS VPN mode can achieve high scalability with the continuous increase of sites and customers. It also supports communication mode between any two points in a site within a VPN without installing a full PVC grid or transmitting traffic backwards on the service provider network. For every MPLS VPN customer, the service provider's network appears to provide a virtual private backbone through which customers can contact other sites within the organization, but not any other client's sites.

From a customer's perspective, an important advantage of the MPLS VPN mode is that in many cases, routing can be greatly simplified compared to PVC mode. Instead of managing routing through a technically complex virtual backbone composed of many PVCs, MPLS VPN customers can use the service provider as the default route to all sites in the company.

VPN service providers often need to provide customers with a range of quality of service (QoS). MPLS VPN utilizes new differential service technology to support QoS. These technologies allow customer traffic to be classified as they enter the provider network based on various strategies such as source sites, application types, etc. Within this network, traffic types are identified by title bits or by different tags, which the router uses to determine queueing treatment and thus determine QoS parameters such as delays and losses.

Traffic engineering

Another benefit MPLS provides service providers is in the field of traffic engineering. The term traffic engineering refers to the ability to control the flow of communications in a network, with the aim of reducing congestion and taking full advantage of available functions.

For example, in this traffic engineering example shown in the figure below, there are two paths from router C to router E. If a router selects one of the shortest paths from C to E, it transmits all scheduled network communications to those that can be reached through E. The traffic brought on this path can thus cause congestion, while the other path is underloaded. To maximize the performance of the entire network, it may be desirable to transfer a portion of traffic from one link to another.

Although one can set the cost of the C-D-E path in this simple example to be equal to the cost of the C-F-G-E path, this load balancing method can become very troublesome in complex topological networks (although this method is not impossible). Display paths implemented using MPLS can be used as a simpler and more flexible way to solve this problem so that a portion of traffic on a crowded path is transferred to a less crowded path.

The solution to the traffic engineering problem is to establish marking and marking switching paths through various control modules. For example, the flow control module may establish a mark exchange path from A to C to D to E, and another path from B to C to F to G to E. The traffic flow on the network can be managed by defining some policies that select certain packets to follow these paths.

MPLS will use restriction-based routing to determine traffic engineering policies in the future. In this environment, simply specifying the expected flow of load between different points of the network (a traffic matrix), the routing system will calculate the optimal path to deliver that load and thus determine the explicit path.

IP and ATM integration

MPLS allows ATM switches to perform almost all functions of IP routers. It has this capability because the forwarding mode of MPLS—mark switching—is exactly the same as the forwarding mode provided by ATM switch hardware. The main difference between a traditional ATM switch and an ATM tag switch is the control software used to establish VCI table entries on the switch. ATM tag switches use IP routing protocol and tag allocation protocol (LDP) to establish such tag entries.

An ATM tag switch can be used as a traditional ATM switch at the same time. In this environment, switching resources (such as VCI space or bandwidth) are divided between the traditional ATM control plane and the MPLS control plane. The MPLS control plane can be used to provide IP-based services, while the ATM control plane provides ATM-oriented services or PVC services such as circuit simulation.

MPLS is a common technology that enables many new services for service providers. It brings a flexible way to leverage QoS (including intranets and external networks) to provide virtual private network services, which provides service providers with good scalability, allows the use of various QoS, and reduces the management burden of VPN customers. The traffic engineering characteristics of MPLS are useful as a way to manage traffic and link utilization of routing networks. Finally, as a method of integrating ATM and IP technologies, it is beneficial for providers who wish to use ATM backbone to establish a multi-service network.

Cisco

Cisco has launched a number of products that provide Tag Switching technology to service providers and large enterprise networks. Current products have the following two basic functions:

  1. In conjunction with an IGP protocol, such as Open Shortest Path First (OSPF), the ability to implement destination-based unicast routing using the Tag Distribution Protocol (TDP);
  2. Travel engineering capabilities to balance traffic loads so as to better utilize all links in the backbone network.
Current MPLS products

Cisco's MPLS products can be used both at the edge and at the center of the MPLS network: an edge device, or an edge switch router (Edge LSR), is a fully functional third-layer device, while a central device, namely a Mark Switch Router (LSR), can be a Layer 2 switch or a Layer 3 router.

Cisco will continue to add MPLS capabilities, such as creating highly scalable virtual private networks (VPNs) that require no tunneling and encryption, providing service level (CoS) capabilities, so that service providers can provide priority to different types of services, even within VPNs. Constraint-based routing capabilities will also enhance traffic engineering capabilities, allowing network managers to optimize the routing of traffic between POPs on the backbone network, while automatically setting up tunnels and mapping traffic to tunnels.

Lucent

Lucent Technology has launched the product IP Navigator based on MPLS technology. IP Navigator and B-STDX, CBX 500 and GX 550 switches can build multi-service solutions, which can provide services based on IP with quality of service assurance in addition to providing ATMs, frame relay services and traditional IP services.

1. Provide Voice Over IP services with absolute service quality

IP Navigator has absolute quality of service function (Absolute QoS), which can establish end-to-end connections that meet voice service quality. Use Lucent Technology's MultiVoice Gateway and IP Navigator to build a Voice Over IP service network with absolute service quality. The MultiVoice gateway provides an interface between the PSTN and an IP-based network, allowing voice calls to enter and exit the IP network. When the voice signal passes through the MultiVoice gateway, it is converted into a VoIP packet voice stream. The MultiVoice gateway sets the service type (ToS) bytes in the voice packet to a specific value and states to the IP Navigator that the IP packet is a voice packet. After the IP Navigator receives the voice packet, it checks whether there is a switching virtual circuit (SVC) between the source and the sink. If an SVC already exists between the source and the sink, the packet will be automatically transmitted through the SVC. If the SVC does not exist, IP Navigator will use end-to-end connections that meet the voice transmission quality requirements according to the requirements to provide voice packets with low latency and low jitter transmission paths, thereby ensuring the service quality of voice. Since ATM is used as the network basis, it can provide large-scale voice communication and is suitable for forming large-scale voice communication networks.

2. IP-based VPN service

IP Navigator's VPN technology can provide multiple VPN networks simultaneously on a MPLS network, supporting multiple applications, such as remote access to local area networks, interconnection of enterprise networks, and Internet applications.

IP Navigator's VPN function assigns a routing domain and an ID number to each VPN user. Each VPN user can have its own independent routing protocol and routing table. The IP addresses used by different VPN users can overlap, so that when users are interconnected through IP Navigator, they can not change their IP address planning. The routing protocols supported by IP Navigator include BGP-4, OSPF, RIP-2 and other routing protocols. IP Navigator has taken security measures on VPNs to ensure that IP connections cannot be established between different VPNs, which is equivalent to physically isolating the two networks. IP Navigator uses the QoS service mechanism of the ATM network to ensure the service quality of VPN users such as bandwidth, transmission quality, and delay. IP Navigator provides user VPN management through user agent gateways, which facilitates users to obtain comprehensive configuration, performance and failure information in real time.

Marconi

(Former FORE Systems Company)

Marconi Communications (FORE Systems) has played a crucial role in the development of MPLS technology for a long time and has provided comprehensive support for MPLS technology.

  1. In a device (such as ASX-4000 backbone switch), MPLS provides support for ATM, POS (Packet over SDH), DWDM and Ethernet at the same time to achieve end-to-end connection-oriented IP. And make full use of Marconi's unique technologies in connection, including TE (traffic engineering), TM (traffic management), CR (capacity-aware routing) and QoS (quality of service).
  2. Implement a connection-oriented link layer (COLL) in one control plane. In the past, IP and ATM integration technology required two sets of control surface protocols. Now, Marconi's MPLS technology only requires one control surface to achieve a complete COLL (Connection Oriented Link Layer).
  3. Provides "Midnight Ship" working mode. Marconi's devices can provide two control plane functions: ATM and IP on the same physical port, so that the same physical network supports multiple control mechanisms at the same time, realizing a smooth transition between ATM and MPLS.
  4. Provide a high-speed and intelligent hardware platform that supports MPLS. Marconi's tag switching hardware provides queuing, buffering, scheduling, and shaping capabilities based on connection and connection groups and ports and port groups. Ensure the best matching of MPLS software and hardware.
  5. Flexible upgrade according to user needs. Marconi provides pure ATM working mode, pure IP working mode and MPLS working mode, and can achieve coexistence and switching of these modes. And all of this is achieved using software upgrades.
  6. Provides large tunnels or connection hierarchical structures. Marconi's MPLS provides a flexible tag stack mechanism, supporting multi-level virtual connections and multi-level VPNs, that is, multi-level virtual connection nesting and VPN nesting.
  7. Solve the IP adjacency problem of N2. Marconi's MPLS uses IGP (Interior Gateway Protocol) to provide direct IP adjacency, avoiding time-consuming and laborious routing update problems.
  8. Simplify the IP table lookup process. Marconi's full-line MPLS products ensure that IP lookup, forwarding and classification are only carried out on connected inbound and exit devices, greatly improving the efficiency of the network.
  9. Supports merging function for connected data streams.

    Marconi Communications (FORE Systems)'s MPLS not only has all the functions required by the IETF standard, but also has very flexible, safe and reliable implementation.

How multi-protocol tag exchange works:

The MPLS specification proposed by the Internet Engineering Task Team places the reliability, delivery mechanism and QoS performance of ATMs on large IP networks. MPLS defines tag-driven connections for reliable and fast IP packet delivery in a multi-vendor environment.

  1. Transfer IP packets to MPLS devices
  2. The MPLS device defines the optimal line for the packet and transmits it on the WAN. At the same time, it also affixes each packet with an MPLS mark. For each hop on the WAN, the router or switch only looks at the MPLS markup and continues to forward the packet.
  3. The MPLS device transfers the packet to another suitable destination, or re-stamps the packet and passes it to a new destination
Article entry: csh     Editor in charge: csh