IP Routing Fundamentals
Author: Mark Sportack
Publisher: Cisco Press (53)
Up to this point, this book has examined many of the underlying technologies
in an internetwork. These have included routers, routing protocols, and
local- and wide-area networking facilities. In this chapter, you will learn
how to integrate these components into an internetwork. Internetworks can
consist of LANs within a single location, or networks that are scattered
across the world in a wide-area network (WAN). Of these two extremes, wide-area
internetworking is the more complex. Therefore, this chapter focuses exclusively
on internetworking via a WAN.
The blueprint for success begins with gathering your users' requirements.
These become the inputs that affect every facet of your internetwork, from
the size and type of transmission technologies, to the placement of routers,
to the choice of routing protocol. Planning a WAN requires the successful
integration of all these technical components.
Successful integration means that the performance of the finished network
meets, or exceeds, performance requirements and user expectations. Therefore,
it is important that you identify and quantify (to the extent that users will
cooperate) these performance criteria before you begin the design. After you
have identified your users' performance expectations, you can begin planning
for your WAN based on several factors:
The first step in planning a WAN is determining its scale. In other
words, how big will it be? Although this may seem to be simple, the word big
can have more than one meaning. In the case of a WAN, scale
refers to the number of locations that need to be interconnected. The greater
the number of locations, the larger the scale of your WAN. Large-scale WANs
can impose interesting challenges for the planner. Scale may prevent you from
using certain topologies, transmission technologies, and even routing protocols!
Therefore, it is imperative that you understand how large your WAN will be
before you start to plan it.
The next factor to evaluate is the distance between
the locations that you are trying to internetwork. Many of the transmission
facilities available today are priced according to distance. This is particularly
true with leased private lines. If your locations are thousands of miles apart,
you might find that the cost of installing a dedicated private line between
them is cost prohibitive. Fortunately, alternatives exist. Many transmission
technologies are priced according to usage rather than mileage. Examples of
these technologies include Frame Relay, X.25, and ATM.
Another way to minimize the cost of transmission facilities is through
careful planning of the WAN's topology. The shape of the WAN can be manipulated
to accommodate the geographic distances separating the locations that are
to be internetworked. Some examples of internetworking topologies are presented
later in this chapter, in the sections titled “Topologies for Simple
Internetworks” and “Topologies for Large Internetworks.”
One of the most important factors to consider when
designing an internetwork is the traffic volumes that it will have to support.
Unfortunately, estimating traffic volumes is an imprecise science. This is
particularly true if there is no preexisting network or communications infrastructure.
If such an infrastructure existed, it would provide some much-needed clues
as to the overall amount of traffic that the new internetwork would have to
Absent such a source of information, your next best bet might be to
interview users to determine the type of work they do, the locations they
need access to, the estimated frequency and duration of their communications
with other locations, and an estimate of the bandwidth that will be consumed
by each communications session.
When collecting your users' usage volumes, keep in mind that there are
actually two traffic volume metrics: maximum traffic volumes and average traffic
In reality, actual traffic volume is almost always volatile; it varies
with times of day, days of the week, business cycles, seasons, and so on.
In other words, you can count on traffic volumes being anything but constant.
Given this volatility, it is important to estimate the maximum amount of traffic
that could be generated at any given point in time. The maximum traffic volume that you expect the
network to support is known as the peak volume.
As its name implies, this is the greatest amount of traffic that you expect
the network to have to support.
If you design a WAN without considering peak traffic levels, it is quite
likely that your users will experience a degradation of performance, if not
an outright service outage, during those times of peak activity.
Average volumes are the traffic loads that you
can reasonably expect during the course of a business day from any given work
location. This type of load is also sometimes referred to as sustained
Establishing these two traffic volumes is critical to the sizing of
the WAN's transmission facilities, as well as its routers. If you expect any
given location to generate an average traffic load of 100 kbps during the
course of a business day, for example, it is clear that a 56 kbps transmission
facility will be inadequate.
Delay is one of the more common metrics that can be
used to measure network performance. Delay is the time that elapses between
two events. In data communications, these two events are typically the transmission
and reception of data. Therefore, delay is the total amount of time that is
required by the network to transport a packet from its point of origin to
its destination. Given this definition, delay is an aggregate phenomenon with
many potential causes. Three of the most common kinds of performance delay
delays are the cumulative amount of time required to transmit, or propagate,
the data across each transmission facility in the network path that it must
take. The size and quantity of each transmission facility in the network path
directly contribute to the aggregate propagation delay of any given transmission.
An additional contributor to propagation delay is traffic volumes. The more
traffic that is flowing across a given facility, the less bandwidth is available
for new transmissions.
Satellite uplink/downlink delaysSome transmission facilities
are satellite based. These require
the signal to be transmitted up to the satellite and transmitted back
down from the satellite. Due to the potentially great distances between
the terrestrial transmission facilities and the satellite, these delays
can be quite noticeable. Uplink/downlink delays are actually a specific
form of propagation delay. For planning purposes, you can safely estimate
the round-trip uplink/downlink
delay time to be approximately half a second. Due to the lengthiness
of this type of delay, however, it is always described separately from
Forwarding delaysForwarding delays in a network are the cumulative amount of time that each
physical device needs to receive, buffer, process, and forward data. The difference
between forwarding delays and propagation delays is that forwarding delay
is measured per device. Propagation delay is measured over the entire network.
Forwarding delay is also known as latency.
Ideally, you would interview your user community to determine their precise
performance requirements, and then select networking technologies and a
topology. Practically, this will never happen. Users tend not to know what
they need. Their ability to assess a network's performance is usually limited
to a reactive, and subjective, opinion. However, users are quite adept at
identifying overall levels of performance that are unacceptable! Acceptable
levels of performance tend to be a function of habit: Normal network performance
is usually defined as whatever they are used to. Obviously, this is a highly
subjective way to assess a network's performance. You may find this description
of a network's possible sources of delays of little help in designing a
WAN. An understanding of them may prove invaluable, however, when trying
to figure out why your users aren't thrilled with
the performance of an existing network.
Designing a WAN ultimately boils down to a balancing
act: The desired performance of the WAN must be reconciled with the cost of
providing that performance level. WAN costs include the initial startup costs
as well as the monthly recurring expenses. Not surprisingly, the larger and
more powerful network components are much more expensive than smaller, less
robust components. The greatest source of expense, however, will be the monthly
recurring charges you incur for the transmission facilities in your network.
Achieving the balance between performance and cost can be painful. No
one wants to design a WAN that will disappoint the users with its performance,
but no one wants to design a WAN that blows the budget either! Fortunately,
several guidelines can help ensure the design of a WAN that satisfies existing
requirements, provides flexibility for future growth, and doesn't exceed the
The capital investments in routers and other network hardware become
a fixed part of the network. After they are placed in operation, they must
be depreciated over three to five years. You might find yourself stuck with
purchased hardware for years after your network outgrows it! Therefore, it
might make more sense to buy a larger router, and then add its internal components
as you need them. This allows future expansion at modest incremental costs,
and little (if any) operational downtime.
Transmission facilities, in comparison to most pieces of internetworking
hardware, are relatively easy to replace with other transmission facilities.
In theory, they can be replaced as often as your lease agreement with the
carrier permits. In practice, swapping out the transmission facilities in
your internetwork can be quite disruptive to operations. At a minimum, you
will need to carefully plan and coordinate every aspect of such an upgrade.
As with physical hardware, your best approach may be overengineering
your transmission facilities relative to projected bandwidth requirements.
Applying the wisdom behind these guidelines can help you meet your users'
present and future expected requirements within the constraints of your budget.
The next step in the planning of your internetwork is the selection of a topology.
Topologies can be highly varied. More importantly, they can be customized
to fit your particular needs. Therefore, consider the sample topologies presented
in the next two sections as information, rather than an actual plan, when
building your internetwork.
One of the primary architects of OpenCable, Michael
Adams, explains the key concepts of this initiative in his book
Broadband, Second Edition
by George Abe
Introduces the topics surrounding high-speed networks
to the home. It is written for anyone seeking a broad-based familiarity
with the issues of residential broadband (RBB) including product
developers, engineers, network designers, business people, professionals
in legal and regulatory positions, and industry analysts.