Chapter 2. ISP Services and Characteristics
This chapter covers the following key topics:
- • ISP Services
A basic categorization of Internet service providers in terms of physical access methods, basic services, and security options.- • ISP Service Pricing
Survey of price ranges for different physical access methods.- • ISP Backbone Selection Criteria
Criteria for evaluating ISPs in terms of their network topology and traffic exchange agreements.- • Demarcation Point
Distinguishing the provider's network, equipment, and responsibilities from those of the customer. - • ISP Service Pricing
Chapter 2
ISP
Services and Characteristics
Before going deeper into the technical subject of interdomain routing, it is important to be familiar with some of the basic provider services and characteristics that affect the quality of Internet connections. Anybody who can offer Internet connectivity could claim to be a service provider; this description covers everything from a provider with a multimillion dollar backbone and infrastructure to a provider with a couple of routers and access servers in his garage.
Price should not be the main factor on which you base your decision. What really matters are factors such as the provider's services, backbone design, fault tolerance, redundancy, stability, bottlenecks, provider/customer equipment arrangement, and so on.
Routing behaviors on the Internet are affected by how routing protocols and data traffic behave over an already established physical connectivity. A good physical infrastructure design and maintenance is one of the main factors in achieving healthy routing on the Internet.
ISP Services
Different ISPs offer different services depending on how big they are and the infrastructure of their networks. Mainly, providers can be categorized by their method of physical Internet access, the applications they provide, and the security services they provide.
The following physical access methods are the most commonly deployed throughout the Internet:
- • Leased lines—Leased lines could be
provided at 19.2 Kbps or 56 Kbps, with increments of 56 Kbps or 64
Kbps up to T1/E1 lines on the lower-end, and T3 and fractional T3
(in multiple T1s) on the higher-end of the bandwidth scale. Leased
lines are usually used when traffic bandwidth is predictable and
the frequency of network access is high enough to justify a line
being up 24 hours a day. Of course, the trade-off is the cost,
which is higher than any other connection type.
- • Frame relay—Frame relay connections are one of the most economical ways for corporations to hook up to the Internet. Purchasing sufficient point-to-point leased line connections can be prohibitively expensive for some companies, in which case they may want to consider connecting to existing frame relay backbones on a per-need basis. With frame relay, corporations can buy enough bandwidth to meet their existing needs and to easily expand as traffic requirements increase. The trade-off is that you are limited by the bandwidth offered by your provider. Other wide area network services such as ATM are starting to be used and recognized, but do not have the success that frame relay has experienced.
- • Dialup services—This includes Asynchronous dialup at 9.6, 14.4, and 28.8 Kbps, ISDN Basic Rate Interface (BRI), or Primary Rate Interface (PRI). Dialup services range from serving individual users to serving corporations that are subcontracting with providers to obtain all their remote login needs. ISDN BRI and PRI services have experienced great growth lately due to their on-demand nature and their capability to carry digital signaling essential for multimedia services.
- • Frame relay—Frame relay connections are one of the most economical ways for corporations to hook up to the Internet. Purchasing sufficient point-to-point leased line connections can be prohibitively expensive for some companies, in which case they may want to consider connecting to existing frame relay backbones on a per-need basis. With frame relay, corporations can buy enough bandwidth to meet their existing needs and to easily expand as traffic requirements increase. The trade-off is that you are limited by the bandwidth offered by your provider. Other wide area network services such as ATM are starting to be used and recognized, but do not have the success that frame relay has experienced.
Prices for ISP services are often predicated on physical access methods, as discussed further in the next section. Customers need to weigh costs and benefits of the different options against their needs.
Almost every single provider offers the following basic services: electronic mail, Usenet newsgroups, ftp, Gopher retrieval, and Wide Area Information Servers (WAIS) resource discovery tool. In addition, due to the phenomenal popularity of the World Wide Web, most ISPs provide this service in one form or another. Customers who do not want to maintain an in-house WWW server can buy space from a shared server on the ISP's premises. Some ISPs are even offering total Web solutions for companies that need a total package, such as advertising, enabling customers to order online, updating inventory, billing, and shipping.
ISPs can offer consulting for different security services. The easiest service would be providing packet filtering on the router level. More security measures involve firewalls and, usually, additional fees. ISPs can integrate their own firewalls or help you configure firewalls that you buy from outside vendors.
ISP Pricing and Technical Characteristics
Besides evaluating services, customers should consider pricing and technical characteristics of an ISP before choosing one. Although technical characteristics in particular may seem intimidating, they have enormous implications for the reliability and ease-of-use of the provider you eventually select. Technical issues that this section addresses include backbone issues and demarcation.
ISP Service Pricing
Prices for services can differ dramatically between ISPs, even for the same services and within the same service areas. The price is usually determined by the provider's relative strength and level of investment in a particular area. A provider, for example, that has an established frame relay service will probably give you a much better price for that service than a provider just setting up frame relay. The following price ranges, based on 1996 data, should give you a rough idea of what to expect:
- • 56 Kbps—First time setup fee is
$1,000 on the average. Monthly fees range between $300 to $1,300,
with an average of $1,000/month.
- • T1—First time setup fee is $1,000 on the average. Monthly fees average $3,000/ month.
- • T3—Prices for T3s vary widely; companies usually strike deals with providers and negotiate the price. The average monthly fee for a T3 link could go as high as $50,000 per month. Customers usually deploy fractional T3 because it is considerably less expensive.
- • Dialup Asynchronous Service—Prices for dialup connections are becoming very inexpensive; prices could range from $10 to $29 per month per user. There is usually a small startup fee and a preset hour limit such as 30 or 60 hours. A small charge is added for additional hours. In some cases, major telephone carriers who are getting into the Internet business are offering free access to their customers, which is driving competitors' prices even lower.
- • ISDN—Pricing is usually provided either as a flat rate between $300 and $500 per month, or based on an hourly rate of about $30 for 30 to 35 hours plus $1.50 to $2 for each additional hour.
- • T1—First time setup fee is $1,000 on the average. Monthly fees average $3,000/ month.
The prices just described do not include the cost of access lines because access lines are provided by carriers such as phone companies. Thus, most ISPs do not include the price of the physical link in their quotes. Customers must negotiate prices of the access lines directly with the carriers, although ISPs usually help to get the ball rolling.
Getting the same price from different providers does not mean that you are getting the same service. Some providers include the Customer Premises Equipment (CPE), such as the router and CSU/DSU (Channel Service Unit/Data Service Unit), as part of the deal. Others charge you extra for the CPE, which could make the bottom line substantially different. You might save a lot of money if you provide your own CPEs and maintain them yourself or pay the provider to do the maintenance.
Large companies usually like to buy their national and international Internet services from the same provider. This means better control and better coordination of services between the different regions of the same network. Some providers offer consolidated billing for all their services, national and international. This means one bill and one check, which is considered a plus by many companies. If the convenience of consolidated billing or common services is not an important issue, companies might find better deals for national and international service, respectively, from different ISPs.
ISP Backbone Selection Criteria
An ISP's backbone encompasses many important technical characteristics. These include physical network topology—connections, network bottlenecks, level of redundancy, distance from destination networks—and traffic exchange agreements. This section is aimed at both customers and designers of ISPs. Customers should certainly evaluate these characteristics in choosing a provider; they are far more important than prices in predicting service quality. Designers should consider the potential benefits and pitfalls associated with these characteristics when setting up or expanding their networks.
Physical Connections
Customers should investigate the provider's physical connections, and the provider should be able to show a decent map of the network, with every connection indicated. With respect to connections, a healthy physical topology is one that can provide consistent, adequate bandwidth for the whole traffic trajectory. The existence of high-speed links such as T3 and OC3/12 does not by itself guarantee overall high-speed access for the customer. Your traffic might flow on the provider's network over some backdoor T1 or frame relay clouds that might slow down the overall access.
Potential ISP Bottlenecks
The provider's network is only as strong as its weakest link. A provider should not oversubscribe its connections. ISPs that save money by overloading their routers or their connections will end up losing customers and credibility in the long run.
Oversubscribing happens when the cumulative traffic of multiple links exceeds the bandwidth of the pipe used to carry the traffic to the rest of the Internet. A provider selling 20 T1s at the POP (Point Of Presence) and connecting to the NAP via a T1 link will experience a bottleneck at the NAP connection. As illustrated in figure 2-1, a rule of thumb is a 5 to 1 ratio: there should be no more than five T1 links for each T1 pipe. Even this ratio might not be accurate if most of the T1s get heavily loaded all at the same time. For ISP network designers, a better approach is to monitor the pipe and to start adding bandwidth whenever line utilization starts exceeding 50 percent of the pipe's capacity.
Figure 2-1 An ISP's
weakest link limits performance.
Another example of a bottleneck is high-speed sites trying to access information from low-speed sites. A Web server located at a site connected to the Internet via a 56 Kbps link can only be accessed at a maximum speed of 56 Kbps, even if you have a faster connection to the Internet. Figure 2-2 illustrates a client llustrates a client with T3 access to the Internet that will be limited to 56 Kbps access to its WWW server.
Figure 2-2 Access speed
is limited by the smallest bandwidth.
If you are a provider, you had better watch the weak links for overutilization, and if you are a customer, understand what you are buying before you commit.
Level of ISP Internet Access Redundancy
Murphy is out there, ready to make your life miserable. Whether because of bad weather, carrier problems, or bad luck, an ISP's connection to a NAP or another POP might go down from time to time, with the potential for leaving customers without service. A redundant network will enable traffic to take another route until the problem is fixed. A well-designed ISP network is a network that has its POPs connected to multiple NAPs or to multiple other POPs, as illustrated in figure 2-3.
Figure 2-3 A redundant
network provides more reliable connectivity.
Distance to Destinations
Customers should be concerned with the number of hops—that is, the number of interim networks—needed to reach a destination network through their ISP. In general, the more hops, the more potential for traffic to get delayed, misdirected, or garbled.
The distance to destinations depends on how the provider is connected to the NAPs. As you know by now, the Internet is a set of overlapping backbones from different providers, with the NAPs being the crossover between one provider and another. Small providers might connect to only one NAP or might not connect to any NAPs at all. In the latter case, a provider could offer service by being the customer of another provider who is connected to the NAP.
In general, providers that claim less than five hops to a destination are those who are connected to most of the NAPs and have a widespread backbone. Smaller resellers tap into major providers, and the traffic might end up hopping a couple of different backbones before reaching the final destination.
Traffic Exchange Agreements
It is important that an ISP be part of all traffic exchange agreements, such as the CIX. Every ISP has to pay its due; there is no free lunch out there. If an ISP is not part of an interexchange agreement with other ISPs, its traffic might be denied when it reaches other providers' networks.
Demarcation Point
Finally, in addition to pricing and backbone issues, customers will want to consider demarcation point (DP) issues in selecting an ISP and forming an agreement. A demarcation point is the point that differentiates the provider's network from the customer's network, as illustrated in figure 2-4. It is important to differentiate between the areas of responsibility of both parties, provider and customer. Demarcation points are defined down to the cables and connectors to make sure that no arguments occur in case of equipment or network problems.
Figure 2-4 Demarcation
point.
Different providers define the DP differently depending on who is paying for the equipment and the access line.
Customer Premises Equipment (CPE)
Customer Premises Equipment (CPE) usually includes the router, the CSU/DSU, the cabling, and probably an analog modem for monitoring. ISPs typically offer customers the choice of buying the CPE and the access line, buying just the access line, or just paying a monthly fee with all equipment and access needs taken care of by the ISP. Any arrangement is available at a price. ISPs usually are responsible for maintaining equipment or packages that they provide. An ISP might have a predefined package that includes CPE and/or access. If the customer does not want to take the package, then the customer would be required to choose equipment that is pre-approved by the ISP. The customer would then be responsible for troubleshooting and maintaining its own equipment. The provider is always available to solve problems at an extra charge. Figures 2-5 through 2-7 illustrate some examples of ISP packages.
Figure 2-5 Example: ISP
provides access and CSU/DSU; customer provides router.
Figure 2-6 Example: ISP
provides access, router, and CSU/DSU.
Figure 2-7 Example:
customer provides everything.
In the scenario illustrated in figure 2-5, the ISP is responsible for the access line and the CSU/DSU all the way up to the CSU's serial connector at the customer site. Restrictions might be imposed on the customer premises routers to meet some memory or software revision guidelines.
In the scenario illustrated in figure 2-6, the ISP has provided everything and its responsibility ends at the LAN port of the router at the customer's premises.
In the scenario illustrated in figure 2-7, the customer provides the CPE and the access line; the provider's responsibility will end at the POP's wiring closet where the ISP interconnects with the carrier's central office (CO).
Router Collocation
Collocation is the act of placing one party's equipment on another party's premises. An example of collocation is putting the customer's router on the provider's site, as illustrated in figure 2-8. The customer motivation for such a collocation scheme would be to have the ISP provide local monitoring of the equipment. Usually ISPs do not like to put customer routers on their premises unless they really have to in order to satisfy the customer. Real estate is becoming a real issue, and the ISPs are having problems finding places for their own equipment at the POP.
Figure 2-8 Example:
customer router located at ISP site.
The opposite of the situation described in figure 2-8 is for the ISP to collocate its own POP router at the customer's site. This is shown in figure 2-9. Usually in this case, the ISP would pay for the access line and the router and would charge the customer a monthly fee for the whole service. The ISP in this case would benefit from more real estate at his premises while providing a total package to the customer.
Figure 2-9 Example: ISP
router located at customer site.
Looking Ahead
Technical characteristics of an ISP's network have significant repercussions for the customer's service, including the quality of routing architecture. Because the customer might not have direct control over some of these technical characteristics, it is critical that the customer at least evaluate them and make sure that they will deliver the required connectivity and quality.
If you are an ISP customer whose demarcation point and collocation agreements stipulate that you are running and maintaining equipment on your premises—even if you do not own it outright—you are likely to be taking a significant hands-on role in developing the routing policies and architecture for your network. Even if you are not running and maintaining the equipment, there are decisions you will need to make and understand with respect to routing architecture.
The next chapter completes the foundational part of this book by discussing addressing and address depletion. After that, all the groundwork will have been established for an in-depth discussion of routing protocols in subsequent chapters.