Module 2 – WAN Technologies
Index
WAN Technologies Overview
The customer connects to the WAN with customer premises equipment (CPE), which is usually either a modem or serial interface. The customer end is known as the DTE.
The DTE from the customer end connects to the DCE, which is the equipment operated by the telco at the other end of the local loop.
WAN standards focus on OSI layers 1 and 2.
The layer 1 standards cover the connection types, such as EIA/TIA-232 and V.35.
The layer 2 standards cover the different types of HDLC used on WAN links. The common HDLC based standards include:
- Cisco HDLC (Point-to-Point)
- PPP (Point-to-Point)
- LAPB (Point-to-Point)
- X.25 (Packet switched)
- Frame Relay (Packet Switched)
- ISDN (Circuit Switched)
WAN encapsulation standards are based on HDLC. At the beginning and end of each frame, there are flags with the sequence 01111110. After every 5 ‘1′ bits, a ‘0′ is added to prevent false ends.
The address field is not needed for point to point links, but is there regardless.
PPP and Cisco HDLC have an extra field – protocol.
Packet switching is used by X.25 and Frame Relay. This routes traffic on a packet-by-packet basis, and allows for more than one logical connection over a single physical line. Charges are based on amount of bandwidth used, not hours connected, which can save a significant amount of money depending on how heavily utilized the WAN link is.
Circuit switching creates a end to end connection, similar to an analog telephone. While these connections allow for a fixed connection between to locations, billing based on distance and connection time can cause for unnecessarily high charges if a link is used on a regular basis with little data throughput.
WAN Technologies
Analog dialup uses the PTSN network to call the remote location and make a connection. There is a maximum of 33kbps, or 56kbps
ISDN uses the wiring for a phone line to make a digital connection to the ISDN switch at the central office. There are two standards of ISDN, BRI and PRI. BRI offers 2 64kbps B channels and 1 16kbps D channel for a total of 144kpbs. PRI offers 23B+D or 30B+D depending on the country, in these situations both the B and D channels are 64kbps.
Leased lines offer a fixed bandwidth between two locations. Speeds vary from 56kbps to about 2.5Gbps, depending on what is available from the provider. Leased lines are a common choice for interbuilding links. The main disadvantage of a leased line is that the link is bought at a fixed capacity, and the bandwidth requirement on WAN links is almost never constant, causing excessive cost.
X.25 was created to allow for WAN links that were billed based on actual data throughput, not maximum speed or link distance. X.25 is packet switched, meaning that several virtual links can run on one physical link. The maximum speed is usually 2Mbps, though providers rarely offered more than 64kbps. X.25 is now uncommon, as Frame Relay has taken its place.
Frame Relay is similar to X.25 in that it is packet switched. The maximum speed is often 4Mbps, with some providers offering higher speeds. The Frame Relay network is usually connected to via a leased line, though some providers offer the ability to dial in via ISDN.
ATM is used when very high bandwidth, low latency connections are required. The maximum speed is in excess of 155Mbps. ATM uses a fixed 53B cell, with each cell containing 5B of headers, and 48B of information. Because of the small cell size, ATM has a very high overhead, typically about 20%.
DSL is a broadband technology that uses the copper local loop from the premises to the CO, where a DSLAM (DSL Access Multiplexer) is used to aggregate several ADSL lines on to a single WAN connection, often T3/DS3. There are several types of DSL, all of which fit into either ADSL or SDSL. Some DSL technologies are able to share a line with telephone services, which means only one physical line is needed to use both phone and ADSL at the same time. Because DSL has to go through the ISP, it is uncommon to use it for WAN links, though VPNs can make DSL a plausible option.
Cable internet connections use the same physical link that provides cable TV. A splitter is used to seperate the TV signals from the data. All the users that connect to the same cable share the bandwidth. All data has to go through the ISP, so VPNs are required to make cable a feasible option for WAN links.
WAN Design
WAN links are usually owned by a communications provider, as it is more cost effective for a large system to be run then leased to users than for users to have to maintain their own system.
WANs carry many types of data, limited only by the capacity of the connections. WANs do not have any services directly connected, as they are for connection of geographically separated LANs.
How to identify and select networking capabilities
The two main considerations to be made when deciding topology are where the traffic is going most and how much traffic. From these considerations, work out which locations need to be interconnected, and how much bandwidth these links will need.
The three layer design groups the link into regions, with links to the central office. This reduces the overall number of links that have to connect to the central office.
Smaller WANs may need need a topology as complex as the three layer model, a two layer model could be used, along with others. Even if the network is small, the three layer model should be considered, as it offers flexible expansion should the network ever need to grow.
Other WAN design considerations
It is possible to use the internet as a way of linking offices in multiple locations, though it is a lot more difficult to keep the network secure with multiple points of entry from the internet, as opposed to having one central connection to the internet. If the traffic volumes are small, the saving of not having to pay for dedicated WAN links could pay for the security required, so it should still be considered as an option.
