.

Unit 3

Kaplan University

IT530
Professor Jeffrey Robinson
August 20, 2013 Abstract
In this discussion we look at some of the history around frame-relay. We then look at some of the equipment necessary to create the frame relay connections, as well as how the connections work both logically and physically. Some limiting factors are discussed as well as committed information rates, burst rates, an error handling. Some of the positive aspects of frame-relay are identified along with some options for the future of WAN technologies.

Frame-Relay
Operating at the physical and datalink layers of the OSI model, frame relay is a high-performing WAN protocol. It was originally designed for use with ISDN integrated services digital network interfacing. It has evolved into being used on multiple types of network interfaces. In 1984 the initial proposal for frame-relay standardization was presented to the CCITT Consultative Committee on international telephone and telegraph. Because interoperability standards were lacking, frame-relay did not fully take off until the 1990s. In 1990 StrataCom, Northern Telecom, Digital Equipment Corporation (DEC), and Cisco created a consortium that focused on the technical development of frame-relay technology. While the CCITT was already discussing frame-relay, the consortium established specifications that while still conforming to the frame-relay protocol, enhanced the protocol with additional features to accommodate more complex internetworking environments. These enhancements are known as the local management interface LMI. On the international front, the international telecommunication Union – telecommunication standards section standardized frame-relay. Frame-relay is an American National Standards Institute ANSI standard. Cisco Kid (2010). There are two categories for devices that are attached to a frame-relay WAN, data terminal equipment DTE data circuit terminating equipment (or DCE). The DTE devices are usually located on the customer’s premises and is the equipment that terminates a specific network. DCE devices are usually owned by the carrier. This equipment provides switching and clocking services in the network which allow data to be transmitted through the WAN. By providing connection oriented datalink layer communications, a defined communication is created between the devices which are linked with a connection identifier. This logical connection between a pair of DTE devices is created with a frame-relay virtual circuit and traverses a frame-relay packet-switched network (PSN). A datalink connection identifier the (DLCI) allow for bidirectional communications between DTE devices in these virtual circuits. By multiplexing multiple virtual circuits into single physical circuits reduces the complexity of the network and the amount of equipment necessary to connect many DTE devices. Inside the frame-relay PSN, one virtual circuit can traverse many intermediate DCE switches. There are two categories of frame-relay virtual circuits: permanent virtual circuits (PVCs) and switched virtual circuits (SVC’s). Switched virtual circuits are connections that are temporary and are created and terminated after each data transfer upon completion of the data transfer. Whereas permanent virtual circuits are connections that are permanent. A DLCI value is allocated to each of the virtual circuits DTE devices within the frame-relay WAN. The virtual connection on each side can have the same value inside the same frame-relay WAN. Cisco Kid (2010).
Some of the LMI enhancements that were developed by the consortium, also known as extensions, provide additional features for complex internetwork management such as multicasting, global addressing, and virtual circuit status messages. Cisco Kid (2010).
Because frame-relay is a technology based on bandwidth demand, the bandwidth is shared on a packet by packet basis with others in the same class. The permanent virtual circuit does not consume any bandwidth until it is needed. This may be cost effective when compared to paying for leased line services. This allows frame-relay to provide a low cost of ownership while adhering to standards and internetworking with other services. A highly reliable WAN may be achieved through frame-relay and a good service level agreement (SLA) from the provider.
Frame-relay can be thought of as a simple principle because it allows the higher-level protocols to deal with many of the problems. Although, it can include optional features such as congestion notification and signaling mechanisms. Ideally, a committed information rate (CIR) is selected that best meets the needs of the customer. This is the maximum burst rate, which when exceeded can cause the frames to be discard eligible (DE). In reality, the nature of frame-relay allows all bits to be discard eligible, but the error correction is handled by the upper layers. Some configurations allow speeds to burst above the CIR, but never more than the physical capacity of the circuit. The DLCI is the connection identifier for the PVC in the connection oriented datalink protocol that is frame-relay. By switching variable length frames, statistical multiplexing is achieved. Even though frame-relay may be susceptible to traffic delays it can still be optimized for delay sensitive traffic such as voice over IP.
Because the frame-relay path is known, it is used to connect the customer premises equipment (CPE) to the frame-relay switch. The LMI keep alive signaling happens between the customer’s router in the local frame relay switch. The frame-relay connects through the PVCs, and the local DLCI’s provide layer 2 identifying connections as assigned by the carrier. Traffic shaping and autosensing LMI are some of the advanced features. Forward explicit congestion notification (FECN) and backward explicit congestion notification (BECN) can control congestion if frame-relay is used inside the cloud of the service provider. D. Harrington (2008).
Primarily, frame-relay is a layer 2 technology that communicates through the provider by DTE device to DTE device, and connections to each datalink segment are connected to the nearest DCE (frame-relay switch). There are two types of encapsulation frame-relay. The default encapsulation type is Cisco, and IETF for non-Cisco device compatibility. The Cisco encapsulation type is used when all devices are Cisco routers, and IETF encapsulation when at least device is not Cisco.
When congestion occurs in the direction from source to destination the FECN notifies the receiving end. Whereas BECN is the opposite and notifies the source. When the CIR is exceeded discard eligible bits may get rejected. These bits are priority based and discards happen when the network lacks adequate resources to handle the traffic beyond the CIR. Because the upper layers request retransmission due to data being discarded, this is why frame-relay can be considered efficient as a layer 2 technology.
LMI provides the signal between the local frame switch and the router. It must match on the same datalink. Request and message between the local router and the frame switch comprise the signaling. D. Harrington (2008).
Similar to ISDN or T-1 connections, frame-relay’s technique is also used as a passage between locations, however there are numerous locations on the network that data can be sent and received to. Each location within the network has connections also referred to as ports. Each port has an address, and each address for a specific location is unique to the port. The equipment that sends data through the frame-relay port sends it in the form of frames, or packets. These frames have two parts, the control block and the actual data, which are sent via virtual connections.
Perhaps an early adaptation of what we now know as the cloud, each frame-relay access port runs inside the cloud which is comprised of many interconnected frame switches. N. Jonker (1997).
Multiple sites within the cloud can communicate with each other through the virtual connections with each connection identified by its unique datalink connection identifier. Unlike traditional methods using dial-up for leased lines, communicating amongst multiple sites that require site to have to dedicated connections which traffic would need to pass through. Frame-relay simplifies this with its virtual connectivity inside a cloud.
This puts the onus of getting the correct data to the correct locations on the equipment. There are two types of virtual connections, fixed and on-demand. Fixed is also known as private virtual circuit (PVC) and is accomplished with data and port addressing via the DLCI. The on-demand type is a temporary circuit that is created when needed. This is called a switched virtual circuit (SVC). Like a phone call, both sides are linked when ready to communicate and disconnected when finished.
Because a single location may handle multiple DLCIs as defined by one port of the frame network, the port is traditionally a T-1 or higher bandwidth physical circuit. And because the multiple DLCIs can technically exceed the port speed frame-relay uses statistical multiplexing to address this issue. This is accomplished by buffering some data in the frame-relay cloud which is later sent as quickly as possible. It is possible however that excessive bandwidth may cause a delay in network traffic and slowness. N. Jonker (1997).
The big difference between frame-relay and other methods revolves around its use of virtual connections as opposed to static connections. Since each location can have a single port into the frame-relay network, the network utilizes multiple virtual connections to its various locations. It adds a layer of redundancy by creating several connections that can be redundant by the use of PVCs between the routers without the need for additional physical links. Because frame-relay does not require a specific media and it can deal with speed differences, it can be a good option for interconnecting different mediums with different speeds with different devices.
Another difference of frame-relay is its burst ability. With traditional connections that are fixed bandwidth, much of the bandwidth goes to waste much of the time because it is not actually always being used. Because the bandwidth on frame-relay is being shared, it allows for sequential multiple bursts which may allow for better bandwidth utilization. However, congestion has a better chance since the frame-relay port has the potential to be a bottleneck. N. Jonker (1997).
One aspect of frame-relay that may make it very attractive is its pricing model. While the price of the port is dependent on its bandwidth, larger ports cost more money, PVC prices are fixed and do not depend on usage or bandwidth. This means that the amount of data sent over the frame-relay when is not subject to extra fees. Probably the best part of frame-relay pricing is that it is not based upon distance at all. When comparing frame-relay options to point-to-point options, one need to only connect a few sites over and intra-lata area for frame-relay to become cost effective. Even locally, when eight or more connections are needed frame-relay again becomes cost competitive to point-to-point circuits. N. Jonker (1997).
Frame-relay is a flexible protocol that can be used for connections of various types it can carry various types of data speed up to 4 Mb per second. Due to the cost is only appropriate for semi-permanent or permanent types of connections, and not a replacement for dial-up services. Frame-relay can also be used to provide Internet services. For customers who have multiple frame-relay ports in different locations, they can leverage this technology to connect to corporate as well as Internet services.
Some of the advantages of frame-relay include offering a high-speed connection over a long distance at less cost than dedicated lines. It allows for the incorporation of redundancy into a network a low-cost. It offers mixed speeds and bursting traffic that can be buffered and converted. Some of the disadvantages include possible congestion and bandwidth is not necessarily guaranteed. For single point-to-point needs frame-relay may not be cost effective. For short distances frame-relay may not be able to compete fiscally with local carriers. N. Jonker (1997).
A WAN is a network of data communications that operates outside the boundaries of a LAN. A LANs three main characteristics are the connection of devices that are distanced, they also use other carrier services such as cell phones, satellites, or cable, and they will typically use serial connections to attach to the wide-area network. So, as their names suggest LANs are for local connectivity inside a smaller geographic region, and may be owned by the institute. WANs are for wide-area connections and likely utilize an external service provider. However, in today’s world it seems that the boundaries between LANs and WANs are blurring. Regardless of the smaller WAN with limited LAN capabilities called a MAN (Metropolitan area network), with the mass interconnectivity of remote users, mobile devices, and Internet capabilities, enter virtual private networks and the MAN/WAN becomes a logical extension of the LAN. LearnCisco.net (nd).
When connecting to WAN service providers it usually happens in the datalink layer and physical layer. The provider handles the operational electrical features of the connections. Options to access the media include frame-relay, ATM, or HDLC encapsulation on serial links. Routers are powerful devices that can offer multiple connections to different services connected via DSU/CSU which converts the signal from the provider. These devices may also identify the demarcation point that indicates where ownership and administration responsibility for the equipment of the provider or customer lies. Some providers offer managed services extend the demarcation point all the way to the DTE device. LearnCisco.net (nd).
But what about network-based applications that are susceptible to packet loss or congestion? Applications like voice over IP, video, RDP, and Citrix may need a more robust WAN infrastructure. When latency and packet loss happen these applications perform poorly, or not at all. Enter MPLS. Similar to frame-relay, ambulances cloud-based. However unlike frame-relay, MPLS can be configured for quality of service (QOS) for WAN applications. Specified applications can be given priority over all of the other traffic on the network during peak times, allowing them to perform adequately while less important traffic is queued up and delivered after the prioritize traffic has been completed. Network Solutions Experts (nd).
There may not be a single best model for WAN technology, but there are numerous options for consumers and companies that when effectively evaluated and designed, allow for optimal communications whether they be data, voice, video, or all at the same time. Regardless if the choice is a dedicated connection, a circuit-switched connection, a packet switched or cell-switched connection, they all have their own advantages and disadvantages. W. Heaton (2007).

References

Cisco Kid (2010). Frame Relay. History of Frame Relay. Retrieved from http://getmydigits.blogspot.com/2010/05/frame-relay.html
D. Comer (2009). Computer Networks and Internets (pp. 329). Upper Saddle River, NJ : Prentice Hall.
D. Harrington (2008). InformIT the trusted tecdhnology learning source. A Brief History of Frame Relay. Retrieved from http://www.informit.com/library/content.aspx?b=CCNP_Studies_Troubleshooting&seqNum=87
LearnCisco.net (nd). Understanding WAN Technologies. Retrieved from http://www.learncisco.net/cisco-courses/icnd-1/wan-connections/wan-technologies.html
N. Jonker (1997). Practical Guide to Frame Relay. Retrieved from http://www.myhome.org/pg/frame.htm
Network Solutions Experts (nd). MPLS Compared with Frame Relay and Internet VPN. Retrieved from http://networksolutionexperts.com/mpls-compared-with-frame-relay-and-internet-vpn/
W. Heaton (2007). TechRepublic/U.S. What's the best WAN connection type for you? Retrieved from http://www.techrepublic.com/article/whats-the-best-wan-connection-type-for-you/