Frame Relay

Outline Why do we need Frame Relay? Frame Relay Network Frame Relay Network as a NBMA network Virtual Circuits DLCI Mapping of DLCI Frame Relay Frame Frame Relay Topologies LMI Frame Relay bandwidth

Why do we need Frame Relay? Frame Relay is more complex a technology than point-to-point WAN links but also provides more features and benefits Let’s take a simple example. Suppose you are working in a big company and your company has just expanded to two new locations. The main site is connected to two branch offices, named Branch 1 & Branch 2 and your boss wants these two branches can communicate with the main site. The most simple solution is to connect them directly (called a leased line) as shown below To connect to these two branches, the main site router, HeadQuarter, requires two serial interfaces which a router can provide. But what happens when the company expands to 10 branches, 50 branches? For each point-to-point line, HeadQuarter needs a separate physical serial interface (and maybe a separate CSU/DSU if it is not integrated into the WAN card). As you can imagine, it will need many routers with many interfaces and lots of rack space for the routers and CSU/DSUs. Maybe we should use another solution for this problem? Luckily, Frame Relay can do it! By using Frame Relay we only need one serial interface at the HeadQuarter to connect to all branches. This is also true when we expand to 10 or 50 branches. Moreover, the cost is much lesser than using leased-lines.

Why do we need Frame Relay? Leased lines provide permanent dedicated capacity and are used extensively for building WANs. Disadvantages A fixed capacity (WAN traffic is often variable) Equipment costs An alternative to dedicated, expensive, leased WAN lines is Frame Relay. Frame Relay is a high-performance WAN protocol that operates at the physical and data link layers of the OSI reference model. Frame Relay provides a cost-efficient solution for communications between multiple remote sites by using a single access circuit from each site to the provider. One disadvantage is that customers pay for leased lines with a fixed capacity. However, WAN traffic is often variable and leaves some of the capacity unused. In addition, each endpoint needs a separate physical interface on the router, which increases equipment costs. Any change to the leased line generally requires a site visit by the carrier personnel.

The capacity between any two sites can vary. Network providers implement Frame Relay to support voice and data traffic between LANs over a WAN. Each end user gets a private line, or leased line, to a Frame Relay node. The Frame Relay network handles the transmission over a frequently changing path transparent to all end users. As shown in the figure, Frame Relay provides a solution to allow the communications between multiple sites using a single access circuit to the provider. The capacity between any two sites can vary.

Frame Relay Network access switches access switches access switches

Frame Relay Network A Frame Relay network is made up of a large number of Frame Relay switches dispersed all over the coverage area of a Frame Relay service provider (e.g. region or country) The switches are interconnected in a complex mesh topology. Frame Relay switches: Terminate user circuits, in addition to connecting to other switches, and are called access switches. Other Frame Relay switches do not terminate user circuits, connecting to other Frame Relay switches only, and make the backbone of the Frame Relay network.

Broadcast MultiAccess networks vs. NBMA network On a broadcast network, multiple devices are attached to a shared network. When one device transmits frames, all nodes on the network "listen" to the frames, but only the node to which the frames are addressed actually receives the frames. Thus, the frames are broadcast. On a nonbroadcast network multiple devices are attached, but data is transmitted directly from one computer to another over a virtual circuit or across a switching fabric. (e.g. ATM , frame relay, and X.25).

Frame Relay Network as a NBMA network Frame Relay networks are multiaccess networks, which means that more than two devices can connect to the network. Unlike with LANs, Frame Relay cannot send a data as a broadcast. Therefore, Frame Relay networks are called nonbroadcast multiaccess (NBMA) networks.

DCE & DTE In the context of Frame Relay: the DTE is the router or sometimes Frame Relay Access Devices (FRADs). The DCE is Frame Relay switch. The purpose of DCE equipment is to provide clocking and switching services in a network + Data terminal equipment (DTE), which is actually the user device and the logical Frame-relay end-system + Data communication equipment (DCE, also called data circuit-terminating equipment), which consists of modem and packet switch In general, the routers are considered DTE, and the Frame Relay switches are DCE. A FRAD is a specialized device designed to provide a connection between a LAN and a Frame Relay WAN.

Frame Relay Network The Frame Relay network is not like the Internet where any two devices connected to the Internet can communicate. In a Frame Relay network, before two routers can exchange information, a virtual circuit between them must be set up ahead of time by the Frame Relay service provider.

Virtual Circuits The connection through a Frame Relay network between two DTEs is a virtual circuit (VC) . VC is a means of transporting data over a packet switched network in such a way that it appears as though there is a dedicated physical layer link between the 2 DTEs. With VCs, Frame Relay shares the bandwidth among multiple users and any single site can communicate with any other single site without using multiple dedicated physical lines.

CH1

Frame Relay is a packet-switched, connection-oriented, WAN service. A VC can pass through any number of intermediate devices (switches) located within the Frame Relay network The circuits are virtual because there is no direct electrical connection from end to end. Provides connection-oriented data link layer communication A logical connection between two DTE across a Frame Relay packet-switched network Provide a bi-directional communications path from one DTE device to another The connection is logical, and data moves from end to end, without a direct electrical circuit. Frame Relay is a packet-switched, connection-oriented, WAN service.

VC Types SVCs are not very common Switched virtual circuits (SVCs): Are temporary connections used in situations requiring only sporadic data transfer between DTE devices across the Frame Relay network. SVC is set up dynamically when needed. SVC connections require call setup and termination for each connection. SVCs are not very common

VC Types Permanent virtual circuits (PVCs): PVCs are more common. A predefined VC. Used for frequent and consistent data transfers between DTEs devices across the Frame Relay network. A PVC can be equated to a leased line in concept. The switching information for a VC is stored in the memory of the switch.

An SVC between the same two DTEs may change. Path may change . A PVC between the same two DTEs will always be the same. Always same Path.

Multiple Virtual Circuits The router or FRAD, connected to the Frame Relay network, may have multiple VCs connecting it to various endpoints. Frame Relay is statistically multiplexed, meaning that it transmits only one frame at a time, but that many logical connections can co-exist on a single physical line.

DLCI Multiple VCs on a single physical line are distinguished by assigning each VC an identifier called data-link connection identifiers (DLCI) . Frame Relay DLCIs have local significance, that means that their values are unique per router, but not necessarily in the other routers. DLCI values typically are assigned by the Frame Relay service provider. Frame Relay DLCIs have local significance, which means that the values themselves are not unique in the Frame Relay WAN. A DLCI identifies a VC to the equipment at an endpoint. A DLCI has no significance beyond the single link. Two devices connected by a VC may use a different DLCI value to refer to the same connection.

Frame Relay Switch Frame Relay creates a VC by storing input-port to output-port mapping in the memory of each switch and thus links one switch to another until a continuous path from one end of the circuit to the other is identified.

DLCI

Frame Relay Address Mapping

Frame Relay Address Mapping Before DLCI can be used to route traffic, it must be associated with the IP address of its remote router

Frame Relay Address Mapping The HeadQuarter will need to map Branch 1 IP address to DLCI 23 & map Branch 2 IP address to DLCI 51. After that it can encapsulate data inside a Frame Relay frame with an appropriate DLCI number and send to the destination.

Mapping of DLCI The mapping of DLCIs to Layer 3 addresses can be handled manually or dynamically. Manually (static): the administrators can statically assign a DLCI to the remote IP address. Dynamic: the router can send an Inverse ARP Request to the other end of the PVC for its Layer 3 address. We should use the “broadcast” keyword here because by default split-horizon will prevent routing updates from being sent back on the same interface it received. For example, if Branch 1 sends an update to HeadQuarter then HeadQuarter can’t send that update to Branch 2 because they are received and sent on the same interface. By using the “broadcast” keyword, we are telling the HeadQuarter to send a copy of any broadcast or multicast packet received on that interface to the virtual circuit specified by the DLCI value in the “frame-relay map” statement. In fact the copied packet will be sent via unicast (not broadcast) so sometimes it is called “pseudo-broadcast”.

Inverse ARP 2 1 Once the router learns from the switch about available PVCs and their corresponding DLCIs, the router can send an Inverse ARP request to the other end of the PVC. (unless statically mapped )

Inverse ARP For each supported and configured protocol on the interface, the router sends an Inverse ARP request for each DLCI. (unless statically mapped) In effect, the Inverse ARP request asks the remote station for its Layer 3 address. At the same time, it provides the remote system with the Layer 3 address of the local system. The return information from the Inverse ARP is then used to build the Frame Relay map.

Now all the routers have a pair of DLCI & IP address of the router at the other end so data can be forwarded to the right destination.

Frame Relay Frame Frame Relay is a data link protocol and the customer router encapsulates each Layer 3 packet inside a Frame Relay frame comprising a header and trailer before it is sent out the access link. The header and trailer used is actually defined by the Link Access Procedure Frame Bearer Services (LAPF) specification The standard LAPF header is too simplistic and does not provide all the fields needed by Frame Relay routers. More specifically, there is no Protocol Type field in LAPF. Each data link layer needs such a field to define the type of Layer 3 packet carried by the data link frame. If Frame Relay uses only LAPF header, routers cannot support multiprotocol traffic because there is no way to identify the type of Layer 3 protocol.

Frame Relay Frame The simple LAPF header was extended to compensate for the absence of a Protocol Type field: There are two Frame Relay encapsulation types: the Cisco encapsulation and the IETF Frame Relay encapsulation, which is in conformance with RFC 1490 and RFC 2427. The former is often used to connect two Cisco routers while the latter is used to connect a Cisco router to a non-Cisco router.

Frame Relay Frame Header Data – Contains encapsulated upper-layer data DLCI , fields related to congestion management Data – Contains encapsulated upper-layer data Trailer Frame Relay provides no error recovery mechanism. It only provides CRC error detection. You should keep in mind that Frame Relay encapsulation should match on the routers at the two ends of a VC. If you fail to match the Frame Relay encapsulation (both sides cisco or both ietf) on the two routers, the connection does not come up. However, if you have Cisco routers at both ends of the connection (a likely scenario), and you don’t explicitly configure Frame Relay encapsulation, both routers default to cisco and the connection does get established. Frame Relay switches do not care about the Frame Relay encapsulation. In Cisco IOS Software configuration, the Cisco proprietarty encapsulation is called cisco while the other one is called ietf.

Frame Relay Topologies When more than two sites must be connected, the Frame Relay topology, or map, of the connections between the sites must be planned. Every network can be viewed as being one of three topology types:

Star Topology (Hub and Spoke) A star topology, also known as a hub and spoke configuration, is the most popular Frame Relay network topology because it is the most cost- effective. Hub Spoke This is the least expensive topology because it requires the fewest PVCs.

Each remote site has an access link to the Frame Relay cloud with a single VC. The hub at has an access link with multiple VCs, one for each remote site. Because Frame Relay costs are not distance-related, the hub does not need to be in the geographical center of the network.

Full-Mesh Topology A full-mesh topology suits a situation in which the services to be accessed are geographically dispersed and highly reliable access to them is required. A full-mesh topology connects every site to every other site More costly than hub and spoke, provides direct connections from each site to all other sites and allows for redundancy Using leased-line interconnections, additional serial interfaces and lines add costs. In this example, ten dedicated lines are required to interconnect each site in a full-mesh topology. Using Frame Relay, a network designer can build multiple connections simply by configuring additional VCs on each existing link. This software upgrade grows the star topology to a full-mesh topology without the expense of additional hardware or dedicated lines. Because VCs use statistical multiplexing, multiple VCs on an access link generally make better use of Frame Relay than single used four VCs on each link

Partial-Mesh Topology Full Mesh Topology Number of Number of Connections PVCs ----------------- -------------- 2 1 4 6 6 15 8 28 10 45 For large networks, a full-mesh topology is seldom affordable because the number of links required increases dramatically. The issue is not with the cost of the hardware, but because there is a theoretical limit of fewer than 1000 VCs per link. In practice, the limit is less than that.

Partial-Mesh Topology For this reason, larger networks generally are configured in a partial- mesh topology. Partial mesh has more interconnections than are required for a star arrangement, but not as many as for a full mesh. The actual pattern depends on the data flow requirements.

LMI Frame Relay bandwidth

LMI The Local Management Interface ( LMI ) is a set of enhancements to the basic Frame Relay specification. It offers a number of features (called extensions) for managing complex internetworks.

LMI Extensions LMI provides a signaling or diagnostic between the DTE router and the Frame Relay switch, using several types of messages: Virtual circuit status messages Global addressing messages Multicasting messages

LMI Extensions VC status messages Provide information about PVC integrity by communicating between devices, periodically reporting the existence of new PVCs and the deletion of already existing PVCs. VC status messages prevent data from being sent into black holes (PVCs that no longer exist). The Status-enquiry messages from the router to the switch sent every 10 seconds allow the router to ask about the status of network. The Status messages from the switch to the router responding to status-enquiry messages. signal if a PVC is active or inactive. VC status messages The Status messages In fact, the Frame Relay switch sends two types of messages: a status message every 10 seconds and a full status message instead of a status message every 60 seconds. contain all the information about known DLCIs and their state Every PVC is predefined by the Frame Relay service provider, but its status can change due to network conditions like failure of trunk links in the provider network. An access link may be up and and running and keepalives may be present but one or more VCs may still be down. The reason is that a VC is an end-to-end logical connection that involves not only the access links at the two ends but also spans the core of the provider network. The router needs to know that which VCs are functional and which are not. The router learns this information as well from the Frame Relay switch through LMI status messages. Every 10 seconds or so, the end device polls the network, either requesting a dumb sequenced response or channel status information. If the network does not respond with the requested information, the user device may consider the connection to be down. When the network responds with a FULL STATUS response, it includes status information about DLCIs that are allocated to that line. The end device can use this information to determine whether the logical connections are able to pass data

LMI status messages act as a keepalive between the DTE and DCE LMI Extensions VC status messages These periodic LMI messages also serve as keepalives for both the router and the switch . If the access link is having a problem, these keepalives will be missed and link problem will be detected. LMI status messages act as a keepalive between the DTE and DCE

LMI Extensions VC status messages LMI status messages combined with Inverse ARP messages allow a router to associate network layer and DCLI. In this example, when R1 connects to the Frame Relay network, it sends an LMI status inquiry message to the network. The network replies with an LMI status message containing details of every VC configured on the access link. Periodically, the router repeats the status inquiry, but subsequent responses include only status changes. After a set number of these abbreviated responses, the network sends a full status message.

LMI Extensions Multicasting messages The multicasting extension defines multicasting as another optional LMI feature. Allows a sender to transmit a single frame that is delivered to multiple recipients. There is a series of four reserved DLCI values (1019 to 1022) that represent multicast groups. The frames sent by a device using one of these reserved DLCIs are replicated by the network and sent to all destinations in the group. The LMI extension for multicasting also defines LMI messages to notify devices of the presence, addition, and deletion of multicast groups. Multicasting supports the efficient delivery of routing protocol messages and address resolution procedures that are typically sent to many destinations simultaneously

LMI Extensions Global addressing messages In the basic Frame Relay specification, DLCI values are locally significant and Frame Relay addresses do not exist. Therefore, mapping must be created. The global addressing extension solves this problem by allowing DLCI values that are globally significant and hence can serve as addresses of individual end routers. The Frame Relay network with global addressing looks much like a LAN to the end routers that can use global addresses (DLCIs) as Frame Relay addresses similar to MAC addresses used in a LAN. The basic Frame Relay specification supports DLCI values that are only locally significant. For example, the DLCI value used on the access link between a router and Frame Relay switch is significant only on the access link and does not in any way identifies the router globally. This DLCI value cannot serve as an address for the router due to its local significance. In other words, Frame Relay addresses do not exist and hence cannot be discovered by usual address resolution methods. Therefore, static maps must be created to tell a router which DLCI to use to reach a remote router. The global addressing extension solves this problem by allowing DLCI values that are globally significant and hence can serve as addresses of individual end routers. The Frame Relay network with global addressing looks much like a LAN to the end routers that can use global addresses (DLCIs) as Frame Relay addresses similar to MAC addresses used in a LAN. Provides connection IDs with global rather than local significance, allowing them to be used to identify a specific interface to the Frame Relay network. Global addressing makes the Frame Relay network resemble a LAN in terms of addressing, and ARPs are used as on a LAN.

LMI Identifier LMI is used to manage Frame Relay links. Each LMI message is classified by a DLCI appearing in the LMI frame.

LMI Identifier The 10-bit DLCI field supports 1,024 VC IDs: 0 to 1,023. The LMI extensions reserve some of these VC IDs, thereby reducing the number of permitted VCs. LMI messages are exchanged between the DTE and DCE using these reserved DLCIs.

LMI Types There are several LMI types, each of which is incompatible with the others. CISCO - Original LMI extension ANSI - Corresponding to the ANSI standard T1.617 Annex D Q933A - Corresponding to the ITU standard Q933 Annex A The LMI type configured on the router must match the type used by the service provider. In order to deliver the first LMI services to customers as soon as possible, vendors and standards committees worked separately to develop and deploy LMI in early Frame Relay implementations. The result is that there are three types of LMI, none of which is compatible with the others. Cisco, StrataCom, Northern Telecom, and Digital Equipment Corporation (Gang of Four) released one type of LMI, while the ANSI and the ITU-T each released their own versions. The LMI type must match between the provider Frame Relay switch and the customer DTE device.

Frame Relay Bandwidth Local Access Rate Committed Access Rate (CAR) Bursting Discard Eligible (DE) Bit

Local Access Rate This is the clock speed or port speed of the the access link or local loop to the Frame Relay cloud. It is the maximum transfer rate at which data travels into or out of the network, regardless of other settings Access rate (AR) is the speed at which the access link is clocked. This is the speed of the physical connection (such as a T1) between your router and the Frame Relay switch

Local Access Rate The service provider provides a serial connection or access link to the Frame Relay network with specific rate (the access rate) These may be 56 kb/s, T1 (1.544 Mb/s), or Fractional T1 (a multiple of 56 kb/s or 64 kb/s). Access rates are clocked on the Frame Relay switch. It is not possible to send data at higher than the access rate.

Committed Access Rate (CAR) CAR defines the bandwidth for the virtual circuit guaranteed by the service provider under normal conditions. The CAR is the amount of data that the network receives from the access circuit. The service provider guarantees that the customer can send data at the CAR. All frames received at or below the CAR are accepted. Originally, when the world was moving from expensive private leased lines to the co-operative model of Frame Relay, customers were concerned about bandwidth because of the contention within the Frame Relay cloud with other customers for available capactity. In order to address these concerns, Frame Relay uses a concept of committed information rate (CIR). Each VC has a CIR, which is a guarantee by the provider that a particular VC would get that much bandwidth. So you can migrate from a private leased line to Frame Relay with a CIR equal to the leased line bandwidth. Committed information rate (CIR) is the speed at which the bits can be sent over a VC, according to the service contract between the Frame Relay service provider and its customer.

Committed Access Rate (CAR) CAR is usually lower than the actual access rate of the interface. Customers can choose the CAR that is most appropriate to their bandwidth needs, as long as the CAR is less than or equal to the local access rate. Typically, the higher the CAR, the higher the cost of service.

Bursting A great advantage of Frame Relay is that any network capacity that is being unused is made available or shared with all customers, usually at no extra charge. This allows customers to burst over their CIR as a bonus.

an access rate on serial port S0/0/1 of router R1 to be 64 kb/s. This is higher than the combined CIRs of the two PVCs. Under normal circumstances, the two PVCs should not transmit more than 32 kb/s and 16 kb/s, respectively. As long as the amount of data the two PVCs send does not exceed the CIR, it should get through the network.

Bursting Because the physical circuits of the Frame Relay network are shared between subscribers, there are often times where there is excess bandwidth available. Frame Relay can allow customers to dynamically access this extra bandwidth and burst over their CIR for free. Bursting allows devices that temporarily need additional bandwidth to borrow it at no extra cost from other devices not using it. Frame Relay allows a customer and provider to agree that under certain circumstances, the customer can “burst” over the CIR.

if PVC 102 is transferring a large file, it could use any of the 16 kb/s not being used by PVC 103. A device can burst up to the access rate and still expect the data to get through. The duration of a burst transmission should be less than three or four seconds.

Discard Eligible (DE) Bit If the customer sends information faster than the CIR on a given DLCI, the network marks some frames with a Discard Eligibility (DE) bit. The network does its best to deliver all packets; however it discards DE packets first if there is congestion.

Frame Relay Bandwidth Several factors determine the rate at which a customer can send data on a Frame Relay network. Foremost in limiting the maximum transmission rate is the capacity of the local loop to the provider. In Frame Relay terminology, the speed of the local loop is called the local access rate. If the local loop is a T1, no more than 1.544 Mbps can be sent.

Frame Relay Bandwidth Providers use the CIR parameter to provision network resources and regulate usage. For example, a company with a T1 connection to the packet-switched network may agree to a CIR of 768 Kbps. This means that the provider guarantees 768 Kbps of bandwidth to the customer’s link at all times.

Frame Relay Bandwidth If the CIR of the customer is less than the local access rate, the customer and provider agree on whether bursting above the CIR is allowed. Since burst traffic is in excess of the CIR, the provider does not guarantee that it will deliver the frames.

Frame Relay Bandwidth An example is a customer buying a 9.6K CIR on a 64K access line. The customer will be guaranteed 9.6K speed but could burst up to 64K if the need arises, for which he may be charged or frames may be dropped.