Using Cisco for Remote Access
As technology increases in importance to businesses, wide area network (WAN) connectivity becomes a primary factor in implementing business practices. WANs are used to connect multiple sites together with a medium that can carry data over longer distance than is possible with local area network (LAN) technology. As businesses expand to new geographic regions, thereby becoming more competitive in the marketplace, the types of connectivity through WANs can differ greatly. In this chapter we will examine the different types of WANs that can be used and the considerations a business (or consultant) must make to effectively choose the right type.
Overview of WAN Connections
Although connecting all users to a LAN is ideal because of the high speed and low cost, it is impractical to do so across large distances. These distances can be as close as an office across the street or as far as an international branch office halfway around the world. These distances require us to examine the types of WANs that are available. For each type, we must keep in mind
Availability: Is the technology available in the area?
Bandwidth: How much do we need, and how much do we get for that type of WAN connection?
Cost: Is there a cheaper connection type that still takes into account future growth?
Ease of management: Is the initial configuration as well as normal operation easy or difficult to maintain?
Quality of Service (QoS): How critical is the actual data itself, and is there a way to ensure low or no data loss with this WAN type?
Security: What measures need to be in place to provide security of company data, while still allowing users and customers to access the data they need?
Reliability: Is the WAN link a critical link, and do we need an additional link in the event of failure?
Application traffic: What is the primary type of data being sent across the WAN, and can this WAN handle that type?
WAN connections can be broken down into three different types, depending on how they carry data: dedicated, circuit-switched, and packet-switched. We will examine each quickly and describe the different types of protocols that each type can provide. We will examine many of these protocols in detail throughout the rest of the book.
Dedicated Connections The first type of connection is a dedicated connection that is used to provide full connectivity between two sites in a point-to-point fashion. Also known as a leased line, this type of connection is purchased from the telephone company (telco) and uses a permanent path through the telco's infrastructure, from one site to another (Figure 1-1). There is no call setup and teardown, which means the circuit is always available.
Figure 1-1 Dedicated connection using CSU/DSU.
Since the company owns the line, it has full use of the bandwidth, whether it is used or not. The speed of the link can range up to a T3, which is approximately 45 Mbps. If the company is underutilizing that bandwidth, then the cost of the dedicated line is high. The cost of the line can also be too great, even if the bandwidth is being properly used, due to distance limitations. As the distance increases and the possibility of crossing geographic (and telco) boundaries appears, the price increases. Therefore, these lines tend to be best used in short distances with a higher volume of traffic or a steady flow of traffic.
This type of connection is usually done with a synchronous serial type of connection. Cisco supports this type with virtually all of their routers, using one or more different types of synchronous serial connections, including
Circuit-Switched Connections There are two types of circuit-switched connections available: asynchronous and ISDN (Integrated Services Digital Network). In both cases the circuit, or dedicated path, is created when the call is initiated to the remote site and the circuit is destroyed when the call ends. The best example of a circuit-switched network is the Public Switched Telephone Network (PSTN) that we use every day in our lives.
Asynchronous circuits for data transfer are accomplished through a modem and the use of the telephone network (Figure 1-2). Since telephones exist in virtually every city in the world, connecting to remote sites is always a possibility. While the cost of the telephone service is very cheap when compared to other types of WAN connections, the real limiting factor is the small bandwidth that is available. Depending on the setup of the connection, the best that can be accomplished is 56 Kbps. Because of this, this type of connection is best used when other WAN types are not available, small amounts of data are exchanged, or cost is a primary issue.
Figure 1-2 Circuit-switched through provider, using modems.
ISDN has two flavors that are used for WAN connections. The first is Basic Rate Interface (BRI) and has a maximum bandwidth of 128 Kbps. This is at least twice as fast as a modem, and the call setup and teardown are much quicker. In addition, having a BRI connection not only allows for data transfer, but analog voice can be used at the same time to cut phone costs to remote sites. BRI tends to be more expensive than asynchronous calls and has limited availability, although ISDN is becoming more available throughout the U.S.
The second type of ISDN is known as Primary Rate Interface (PRI) and can reach speeds up to 2 Mbps. This type of WAN connection is ideal for combining multiple BRI channels and asynchronous calls into the same router, using only one physical interface.
There is one other potential use for both asynchronous and BRI networks. In the event that a primary link fails, they may act as a backup link to ensure connectivity.
Packet-Switched Connections Packet-switching (Figure 1-3) is a method where two or more sites are connected through a shared network, typically called a cloud. By shared network, we mean that more than one company has access to the cloud. Remote sites are connected via a virtual circuit (VC) that allows data to traverse the cloud and arrive at the correct location. Within the cloud, each packet can take a different path to reach the final destination. Because the data travels through a shared cloud, the cost tends to be lower than the same bandwidth used for a dedicated line.
Figure 1-3 Packet switching using provider cloud such as Frame relay.
Packet switching can be considered the common ground between dedicated lines and circuit-switched. Although usually more expensive and not as freely available as circuit-switched networks, the additional bandwidth (up to T1 speeds) makes it an attractive alternative. Also, it is cheaper over longer distances than dedicated lines, which again makes it a nice alternative.
Protocols Used on WANs
There are many different types of protocols used throughout the world, depending on location, type of WAN, and administrator knowledge. These protocols all operate at layer 2 (at least) of the OSI model (data-link layer). We will briefly review them here, and other chapters will dedicate more discussion to the more common of them.
Point-to-Point Protocol (PPP) PPP is used with both dedicated lines and circuit-switched networks. It is a standard protocol that vendors can use to interoperate their equipment with other vendors. In addition, PPP supports multiple network layer protocols such as TCP/IP and IPX/SPX, as well as authentication and compression mechanisms. Because PPP is such a versatile protocol, we will examine it more thoroughly in later chapters.
Serial Line Internet Protocol (SLIP) One of the earliest protocols used in point-to-point connections, SLIP is being phased out due to some of its serious drawbacks. These drawbacks include support only for TCP/IP and lack of security. SLIP can be used over dedicated and circuit-switched networks. We will not examine SLIP in this book.
High-Level Data Link Control (HDLC) Although HDLC is a standard, the limiting factor of that standard is the support for only a single protocol. Because of this, Cisco has modified it to support multiple protocols over point-to-point links. This is the default encapsulation protocol on serial links with Cisco routers. Because of this modification, though, Cisco products may not interoperate with other vendors' equipment. In this situation, PPP is the better protocol. HDLC is supported over dedicated lines, but not over circuit-switched or packet-switched networks. HDLC is a simple protocol, and therefore we will not examine it in much more detail in this book.
Frame Relay (FR) Frame Relay was a protocol derived from ISDN specifications and has evolved to become the dominant layer 2 protocol over packet-switched networks. As a standard, it has become widely available throughout the United States and many other parts of the world. Frame Relay is a protocol that should be thoroughly understood, and so we will examine it in later chapters.
X.25 X.25 protocol is an older protocol that can still be found throughout the world. It is also a protocol that is used over packet-switched networks. The primary difference between X.25 and Frame Relay is the overhead of error correction built into X.25. This overhead was a necessary feature of the older, less reliable networks and networks still found in more remote parts of the world. Although it is an older protocol, it does have a large install base worldwide, and therefore we will examine it in more detail in a later chapter.
Future Networks Asynchronous Transfer Mode (ATM) networks are cell-switched instead of packet-switched. This smaller, fixed cell size of 53 bytes allows multiplexing data such as voice and video with more control. ATM is designed for very high speeds of data transfer, such as OC-192 (10 Gbps!).
Digital Subscriber Line (DSL) and cable modem technologies are other types of WAN connections that are being brought to the public's attention. Their primary role so far has been to connect more people to the Internet using faster speeds than is capable with a modem.