Chapter 3

Knowing Your Networking

Chances are that you are knowledgeable about some aspects of networking, but are somewhat uncertain about others. Rather than make an assumption that you are a universal expert on all things networking, this chapter covers the most important technical elements of networking. Networking expertise is critical to your success in a networking job.

A great deal of time in this role is spent planning, budgeting, presenting, training, and generally collaborating with coworkers. We deftly avoid mentioning these administrative aspects of networking as of the end of this sentence.

Appreciating First Principles

Before we get started, a few things need to be said about the network in an organization. First we cover some basic concepts about the design goals of a network. Next we describe business objectives and how these affect network design.

Recognizing design goals for the network

There is no such thing as a one-size-fits-all network. Each organization has unique requirements:

  • Connection to legacy computer systems: Unless you work at a startup launched just minutes ago, your organization has existing computer systems for things such as email. Connecting your computers together (and to the Internet) involves a network.
  • Plans for future systems: Organizations are dynamic and always evolving, including adding applications to increase customer service or reduce costs. New applications inevitably involve changes to the network.
  • The locations of offices: Many organizations have remote offices that need connectivity to computer systems. Companies are regularly adding and reconfiguring remote offices.
  • Mobility needs of employees: Companies have been issuing laptops to employees for years. Employees are now using tablets and smartphones, (their own or ones supplied by the company) to access company systems.
  • Throughput: As IT offers new services, users need more capacity. The good news is that high-speed WAN service is getting cheaper (although it is still much more than LAN service).
  • Internet connectivity needs: In addition to web hosting, many companies need to provide Internet connectivity for business use. It may or may not be company policy to allow the personal use of Internet during business hours.
  • Redundancy and resiliency: The architects of the network need to decide which approach is the best fit for ensuring that business operations will continue when the inevitable failures happen on and around the network. A range of solutions keeps a network up and running. Choosing among redundancy options is typically limited by budget and also by the personnel involved in restoring a failing network when all heck breaks loose.
  • Level of security: Just like redundancy and resiliency, you can pile on security until it becomes excessively costly and difficult to manage. The company needs to decide what level of security it wants to employ to achieve its overall business objectives.

Appreciating the business objectives of the network

It would be pure fantasy, and probably a bad idea, to think that a company would consider limiting its growth and business plans to meet the existing design of its computer network. The computer network, in all its capabilities and restrictions, needs to work around the business.

This concept of the network serving the business entails more than just coming in under budget during the current fiscal year and more than accepting budget cuts to achieve profitability during business planning. The entire computer networking team needs to be on board to provide remedies to the inevitable network glitches that occur.

Proper network management includes planning for server, LAN, WAN, and power outages; PC viruses and other malware; security breaches; and lightning strikes and other severe weather. The entire department needs to know what to fix first, second, and so on. Proper planning and operational management is not something that should be left to leadership skills implemented in real time.

Reviewing the OSI Model

It is an unwritten rule that every discussion of networking should include the 7-layer OSI (Open Systems Interconnection) model, depicted in Figure 3-1.

Image

Figure 3-1: The OSI 7-layer networking model.

The OSI model breaks the network communications process into seven separate layers. From the top (the layer closest to the user) down, these layers are as follows:

  • Layer 7, Application: The Application layer provides services to the software through which the user requests connectivity. This layer is about programs such as browsers, FTP clients, and mail clients.
  • Layer 6, Presentation: The Presentation layer is concerned with data representation and code formatting.
  • Layer 5, Session: The Session layer establishes, maintains, and manages the communication session between computers.
  • Layer 4, Transport: The functions defined in the Transport layer take care of reliable transmission of data segments, as well as the disassembly and assembly of the data before and after transmission.
  • Layer 3, Network: The Network layer is where routing takes place. As a result, this layer is perhaps the most important OSI layer for to know. We may even rename this book OSI Model Level 3 Jobs For Dummies (or not).
  • Layer 2, Data Link: The Data Link layer is concerned with the mechanisms used to move data about the network, such as Ethernet, and deals with the ways in which data is reliably transmitted.
  • Layer 1, Physical: The Physical layer defines the electrical and physical specifications for the networking media that carry the data bits across a network. Think of this layer as referring to copper, fiber optic, and wireless.

The 7-layer OSI model is imperfect. Many manufacturers develop products that touch on multiple layers. However, nothing beats this model for communicating a conceptual understanding of networks.

Exalting TCP/IP

Almost as pervasive as the OSI 7-layer model is the use of TCP/IP in networking equipment. It has become the universal solvent for all problems networking. Developed in the 1970s as a robust military communications network that had some self-healing properties and resilience, TCP/IP has formed the basis for virtually every home, business, and commercial network, as well as the global Internet itself.

TCP/IP is a packet-based technology in which messages are bundled into packets that are routed to their destinations. A single packet has a source address, a destination address, a protocol number, and a payload (the contents of a message). Figure 3-2 shows how TCP/IP maps to the OSI 7-layer model.

Image

Figure 3-2: The OSI seven-layer networking model mapped to TCP/IP.

Staying Local with LANs

As long as you stay within your own property, you can network using LANs. As a rule, LAN technology is much cheaper and faster than WAN technology. An organization needs to use WAN technology when it must connect two LANs across some public right-of-way such as a street or a river.

The largest property under one fence in the United States is the Waggoner Ranch in northwest Texas. Ignoring propagation issues, the half million acres (about two-thirds the size of Rhode Island) could all be on one LAN. The instant they want to communicate with the bunkhouse in nearby Vernon, TX, however, they need to use WAN technology. That is just the way things work.

Using wired LANs

With some exceptions we discuss later in the chapter, pretty much every wired LAN is a variation of Ethernet. Four kinds of hardware elements are part of Ethernet data networking:

  • Network interface card (NIC): This card goes into a PC, laptop, or smartphone.
  • Hub: The hub connects Ethernet cables from multiple devices so that they can all communicate. This simple, unintelligent connection option is economical, particularly for a small office or a home network.
  • Switch: An Ethernet switch combines a hub with a little bit of intelligence. Whereas the hub has each NIC card listen to each message and determine whether or not the packet is meant for it, an Ethernet switch will figure out which PC the message is for and not bother every NIC card with every message.
  • Router: The term router applies to a range of equipment to manage an Ethernet LAN and interact with WANs. These devices range in complexity from routers in a small home office up to carrier-class devices that control the universe (or at least the digital part of the universe).

Image Some manufacturers incorporate a wireless LAN access point into a router. Other manufacturers don't. It is easy to be confused.

Ethernet technology has the lion's share of the installed base of LANs. The different flavors of Ethernet are based on cable and nominal speed, as shown in Table 3-1.

Table 3-1 Ethernet Speeds

Image

Although Table 3-1 looks straightforward, there are several opportunities for confusion. First, some people refer to the 10Mb per second local area network as Ethernet. Others refer to the technology that includes all the options in the table as Ethernet. Both are correct, but the terminology can be confusing.

To distinguish among them, many people refer to the cable connector, such as 10BASE-T. However, the cable connector is also called RJ-45. On top of that, the cable can be called Cat-5 or Cat-6 cable. The best advice is to ask for clarification on the Ethernet speed only if it matters.

Another source of confusion is the speed of the Ethernet. There are two reasons why Ethernet speed issues can get confusing. A cool feature of this technology is that it will “dumb itself down” to accommodate the slowest technology. For example, your router and wiring may accommodate Gigabit Ethernet. However, if the NIC can work with only 10Mb per second, the technology will operate at the slower speed. The problem is that the technology will never tell you that it is operating below what it's capable of. You have to know to ask. The 10/100/1000 specification indicates that the equipment is happy to work at any of those speeds and will default to the highest possible.

Another side of Ethernet speed is that 10Mb per second is the nominal speed. TCP/IP is chatty. By chatty, we mean that the respective ends of the connection spend a fair amount of time making sure that the other end is ready to receive the transmission, that it received the transmission correctly, and that it has finished sending. These tasks are important to ensuring an accurate transmission, but they require bandwidth. You, as a user, get only a percentage of the nominal speed. This topic is described in the next section.

Using non-Ethernet LANs

Ethernet LANs are the most common type of LAN. The hardware is inexpensive and readily available. Plus, many people know how to support it.

There are many options for non-Ethernet LANs. They are typically used in special applications, such as when exceptional security is required.

Token-ring LANs, for example, were heavily promoted by IBM in the 1980s. Eventually, IBM gave up and accepted that Ethernet had the dominant market share. As the name implies, a token-ring LAN has all PCs on a logical ring. The PCs on a given LAN ring are in communication with the PC that is logically to its left and to its right. When it receives the token from the PC to its right, it looks to see if that token has a message that belongs to it. If it does, it takes that message and hands off the token to the PC on its left. If the token does not have a message for that PC, it passes the token along with the message to the next PC on the left. This process happens very fast.

This architecture feels more reliable compared to the Ethernet, where all the PCs essentially repeat until a message is acknowledged.

Another alternative to Ethernet is Fiber Distributed Data Interface (FDDI). This technology operates at a nominal rate of 100 Mbit/s. Whereas Ethernet uses copper wire, FDDI employs fiber-optic cabling. In addition, each PC is connected back to the central router in a star topology. These two features make FDDI more secure than Ethernet. (Note, however, that some Ethernet LANs use fiber-optic cable.)

Moving around with wireless LANs

The benefits of wireless LANs have been understood for quite a while. Running cable to offices through walls is annoying. Also, workers want to take their laptops and smartphones to meetings and still get emails and access to applications.

Wireless LANs did not gain popularity until the last decade or so. The first standard that caught on was 802.11, announced in 1997 by the Institute of Electrical and Electronics Engineers (IEEE) workgroup on wireless LANs.

Image The common term for 802.11 wireless LAN is Wi-Fi. This term is licensed by the Wi-Fi Alliance. The Wi-Fi Alliance picks up where the technical specifications leave off by providing a certification process to ensure that elements from competitive manufacturers work together.

This workgroup has since come out with new specifications that serve a wider area and add new frequency bands and, more importantly, higher data rates. The specifications and their nominal data rates are shown in Table 3-2.

Table 3-2 802.11 Specifications

Image

Note several things in this table. First, as with Ethernet, the elements in the system automatically adjust to the fastest protocol. You may put in an 802.11ac system, but if the laptop operates at only 802.11a, the protocol uses is 802.11a, which will limit the speed. Unless you check, you never know the speed at which a given laptop is connecting.

Also, the speed mentioned is at optimal conditions. Your mileage may vary — a lot. Wireless communications are fickle; one moment you may be crawling down the highway and the next you may be screaming down it.

The distance from the antenna is a big factor in the speed that a user will see. Table 3-2 shows the maximum distance at which the given protocol is designed to operate if there is a line of sight between the antenna and the device. This value is approximate and changes in real time. One solution for offering more complete coverage and avoiding dead spots is to put out multiple access points to serve a particular office space. Multiple access points work up to a point. It would take an entire Dummies book to cover all options and strategies for addressing coverage with wireless LANs.

Although wireless LAN coverage is important, security is more of an issue. Wireless LANs might seem like a security problem waiting to happen. However, a properly implemented wireless LAN can be more secure than some copper wire LANs. The key phrase in that last sentence is properly implemented.

A wireless LAN in an organization as well as in a home should be password-protected. Leaving a wireless LAN unprotected is an invitation for trouble. Some bad people make a habit out of walking around looking for unprotected LANs. They may just use some of your service when you want to use it. More likely, these will try to access your LAN and steal credit card and bank account numbers. Most systems allow you to enter a password once to access Wi-Fi. The password is then stored on your PC or smartphone for when you want to get on the LAN in the future.

Going Out with WANs

As mentioned in the section on LANs, the minute users step off the Waggoner Ranch onto route 183, they must connect to a WAN and not a LAN if they want to connect to the computers on the ranch. Historically, the easiest way to connect from here to there has been a dial-up circuit over telephone lines using a modem.

The first modem, the Bell 101, transmitted data at 110 bits per second. This speed was far superior to what a skilled Morse code operator could deliver (20 bps) and was comparable to what a skilled typist could key in a Teletype machine (50 bps).

Even better, the pair of Bell 101 modems would also be on duty at all times, never needing a vacation or a break. For an investment of a few thousand dollars, you could avoid paying a human operator $2 per hour.

Modems added intelligence over the years, so that from the early 1960s to the late 1990s, they could eke out up to 56kbps from a single phone line. This is the greatest speed you can practically get from a single phone line, called a DS-0 in telephone lingo.

Speeds available on a DS-0 and below are called narrowband. Anything above DS-0 is broadband. These days, we think of narrowband as quaint. Broadband signals available from service carriers include DS-1, T-1, E-1, DSL, SONET, DOCSIS, MPLS, and dark fiber. (For details on these, see Chapter 3.) Older technologies you don't need to be too concerned with anymore (unless you're a technology history buff) include ISDN, ATM, Frame Relay, X.25, and PSTN.

Buying service from LECs and CLECs

Many data connection services are available from your local phone company. The “local phone company” refers to the company that used to have a monopoly in your particular area. Until the last few decades, some telecommunications companies were given exclusive rights to provide phone and other WAN services in a particular region, such as a city or a specified unincorporated area.

The deal, forced primarily by the Federal Communications Commission (FCC), was that local phone companies would provide local voice telephone service, and other companies would provide long-distance services. The local phone company in your area was called a local exchange company (LEC).

To make a long story short, the LECs wanted to get into providing longdistance voice service. The deal struck enabled existing LECs to offer longdistance services if they allowed other companies to provide local phone service. These were called competitive LECs (CLECs).

Some CLECs ran copper wires to provide service in areas with significant new construction. This practice was the exception rather than the rule. For the most part, CLECs carried high-capacity fiber-optic cable and would provide services to large building and corporate campuses.

Considering the cloud and managed service

An organization can buy individual links to connect the main office to a remote office. These can come from either a LEC or a CLEC. Another option is to buy networking services from a company that specializes in this area.

It is common practice to contract with an Internet service provider to provide Internet connectivity for organizational use and for web hosting. The extreme case of outsourcing network services involves relying upon the cloud. At this point, the only network access you need is a large pipe for Internet access. The cloud provider manages all server storage and security access. The use of cloud-based services is a viable strategy for startups.

Many LECs and CLECs offer managed networking services for their clients. Several IT companies, such as IBM, HP, and Level 3 Communications, provide this kind of service for organizations.

Accessing Wireless WANs

Unlike wired service, wireless WAN access is available from many sources. The largest providers these days are cellular carriers. In addition, you can get your own microwave connection, satellite service, or private radio.

Going beyond texting on a cellphone

Not too long ago, cellular carriers realized that you could send data over their cellular networks. A few pioneering souls back in the mid-1990s connected special modems to their cellphones. The cellular carriers found that these customers used three or four times as much airtime, never called for customer service, and did not switch their service to save a few pennies.

As a result, cellular carriers went on a competitive binge to offer faster service in more places. These new offerings were great because we all received faster service at a lower cost. However, all these new names and claims were confusing, with all kinds of acronyms and marketing hyperbole.

The different cellular companies would all claim that their network was the best and the fastest. They would claim that their network was 1G, or 2G, or 3G. They never explained that 1G meant first generation, 2G meant second generation, and so on.

They were also comparing apples and oranges. Sprint, Verizon, and US Cellular use a digital technology for voice called CDMA. AT&T and T-Mobile use a digital technology for voice call GSM. 2G technology on CDMA is different than the 2G technology on GSM.

Moreover, whether the user is stationary, moving (as in walking), or driving makes a significant difference in data throughput. This factor was conveniently left out of the advertising copy. The Radio Section of the International Telecommunications Union (ITU) says that 3G should be up to 7.2 Mbps. They also said that 4G should be 1Gbps when stationary and 100Mbps when moving.

However, ITU does not officially own the term 4G, plus other requirements are not met. The bottom line is that cellular carriers can call whatever they want 3G or 4G and get away with it.

What is more important is whether or not you are in coverage. The concept of coverage is easy to understand when dealing with voice service but less obvious when we are talking about data.

Image Ignore whatever G the carrier is telling you they have and pay attention to their coverage.

Cooking with a microwave connection

No, we aren't talking about cooking in the literal sense. We are talking about microwave communications. If you have to connect to buildings across the street, you may be surprised how economical and convenient it is to set up a point-to-point microwave system. As long as you have line-of-sight from point A to point B, many companies will put up microwave systems that are more economical over the long run than buying ongoing service from a LEC or CLEC.

Microwave is a telecommunications technology that has been used since World War II. Telecommunications companies have provided service between switching center offices for decades. About one third of the millions of miles of lines managed by LECs are provided by microwave.

Point-to-point microwave communication has several gotchas. First, you have to have line of sight from point A to point B. Next, you need access to the roof, which can be complicated if you are a renter, or you need to “shoot” through windows, which reduces your range. If these are not big issues, you should consider microwave as an alternative or a complement to service from the LEC.

Another option for microwave communication is WiMAX (Worldwide Interoperability for Microwave Access), which is a technology used by the Clearwire network.

WiMAX is a wireless telecom standard that provides data rates up to 40Mbps for mobile stations and 1Gbps for fixed stations. WiMAX was developed to be a wireless alternative to DSL and DOCSIS.

Clearwire offer service in 17 markets in the United States and is targeted to small businesses and fixed wireless applications.

Circling around satellite communications

Satellite communications seemed exotic at one time. These days, almost anyone can acquire television service from DirecTV or Dish Network and install his or her own satellite receiver.

The data-only equivalent satellite TV service is called VSAT, or Very Small Aperture Terminal. Depending on the service you acquire, you can get uploads and downloads in the narrowband or broadband range.

There are a few important considerations when considering VSAT service. First, VSAT terminals have a longer latency than ground-based systems. Regardless of the volume of data and the rate of speed on the network, it takes a few seconds to get everything flowing. This factor is not an issue for some applications and a showstopper for others. If you want to send and receive large files, the delay is not a problem. If you're doing lots of short transactions, the delays may drive users batty.

Another consideration is that VSATs are good at sending large files to a remote site, but are much more limited sending files from the remote site back up to the satellite. Again, slow upload speed is either a big problem or no problem depending on your circumstances.

VSAT excels in providing a backup to landlines. Landline service can fail, for example, when a backhoe rips up communications cables. A backhoe can't rip up communications between a satellite and the ground station antennae.

..................Content has been hidden....................

You can't read the all page of ebook, please click here login for view all page.
Reset