Designing Network Architecture

If you wanted to build a house, you wouldn't just go to the lumber store and the hardware store, buy some supplies, and begin construction, without a blueprint of what you wanted to build. Too many things need to be taken into consideration: rooms, doors, windows; water pipes and electric cables; outlets, and so on. These different elements must be visualized and thought through ahead of time, unless you have infinite money and infinite time, and don't really care about what your house ultimately looks and feels like.

• Designing the Network Achitecture: You need the network architecture document for the same reasons you need a blueprint to build a house. This document defines the overall integration of the various hardware and software components used by an organization to develop their Intranet, Extranet, and Internet sites. As a visualization tool, the network architecture document is useful in ensuring that the needs of the site and the resources available for it are understood by the team as well as the client.
• The Web Team: For the Web site development team, the primary concern is the ability of the network to support the business objectives of the Web site. IT staff, or in the case of Webteam, the Technical team, generally make the key decisions regarding the network architecture. However, the Web site development team should ensure that the IT staff is aware of the technological needs that the site will require to function smoothly.
• The use of the Requirements Definition: The main source of information when designing the network architecture for a site is the Requirements Definition. As you know, the Requirements Definition lists the hardware and software requirements, as well as the type of connections and other specifications related to the hardware to achieve the site's business objectives. To go back to the house analogy, the Requirements Definition lists the materials needed, while the network architecture illustrates how these materials will be used and integrated into a whole.

The diagram below illustrates how a sample network architecture is put together. It is much easier to understand the configuration needed for a given project when you can see the relationship between the different parts.

What will happen if various pieces of the network hardware crash? What provisions exist for backing up web files? For backing up customer and transaction data? Network crashes have various flavors. A router effects a segment, while a server could truly bring the whole network down. It all depends on redundancy and criticality of the hardware's location. Click the sidebar to learn more about keeping your system failsafe.

Backup of data is such a primary issue that gross negligence is now assumed whenever server-based data is lost. RAID (dual mirrored drives) are the norm in such environments, and use of tape is still a nightly routine (or even four times per day). I work in an evironment where ALL server-based user data is backed-up every five minutes!! Finally, mission critical data, such as Am-Ex, or VISA transactions from around the world are continually stored in huge repositories six-stories underground in Texas to assure that, in the event of a thermo-nuclear exchange, you would still get your CC bill the moment the rubble was cleared away
What will happen if various pieces of the network lines go down?
The implications of an outage are dependent on the criticality of the line. It may mean no email for while, or it may inhibit department-wide process activities if data is sourced from other areas, such as call centers, as opposed to fulfillment centers.

• What is RAID? RAID (Redundant Array of Inexpensive Disks) is a data storage structure that allows a system administrator, designer, builder, or user to combine two or more physical storage devices (HDDs, SSDs, or both) into a logical unit (an array) that is seen by the attached system as a single drive.
  There are three basic RAID elements:
  1. Striping (RAID 0) writes some data to one drive and some data to another, minimizing read and write access times and improving I/O performance.
  2. Mirroring (RAID 1) replicates data on two drives, preventing loss of data in the event of a drive failure.
  3. Parity (RAID 5 and 6) provides fault tolerance by examining the data on two drives and storing the results on a third. When a failed drive is replaced, the lost data is rebuilt from the remaining drives.
  It is possible to configure these RAID levels into combination levels, called RAID 10, 50 and 60. The RAID controller handles the combining of drives into these different configurations to maximize performance, capacity, redundancy (safety) and cost to suit the user needs.

In the next lesson, you will learn about how bandwidth is measured and how bandwidth needs are evaluated.

There are four basic elements of network architecture that must be considered when designing a network. Each of them must be planned when designing the network. These elements are illustrated below:

1. Fault tolerance: the mean of fault tolerance is to ensure that the network will function properly even if fails occurs. The fails can occur in many different locations inside the network such as cables, routers, switches, or servers.
   The key idea in fault tolerance is to make duplication for the above components such that if as fault occurred in each of them, there will be an alternate one and the network will still be functional. This is a popular component in the internet architecture design.
2. Scalability: the meaning of scalability is to ensure that the network can grow in the future without changing the design. Thus the administrator of the network can add more users to the network or add even an entire sub network without rebuilding the original network. This element is very important on the internet because any one can see that the internet is growing continuously and more users are added every day.
3. Quality of service (QOS): QOS is also an important aspect in any network especially on the internet. QOS enables the admin to classify the services performed by the network from higher to lower. Thus the network performs different from one user to another according to the data he is sending or the type of the user. Some users transmit voice on the network when making a call while others transmits regular data such as files and emails. Each one of these services requires a different network usage or a class of service. The voice and video must take the highest priority while the files and emails take the lowest priority. This is because voice and videos are delay sensitive and must be transmitted as soon as possible.
4. Security: when transmitting sensitive data across any network such as the internet, security often is an important aspect to consider. This is because the data to be transmitted may be sensitive or private. This is solved by encrypting the data such that no one can read the data other than the destination. On the internet, a protocol called secure hypertext transfer protocol is used for this purpose.
   The sensitive data may be passwords or credit card numbers which needs to be protected. Also when logging into a server, user credentials needs to be defined first at the server. This is called authentication and is another element of security. The two are very important aspects that are used in the internet protocols. Note that the internet is an open network and so security is very important as a part of its architecture.

SONET links can be thought of as highways, as SONET is used in metropolitan and wide area networks. Using this highway analogy, we can say that a SONET transport network is hierarchical, as a collection of small roads, medium-sized roads, and highways for long distance. Unlike highway systems, the typical architecture for SONET is a collection of rings, although point-to-point connections are used as well.
Figure 4-4 illustrates the following three classes of SONET networking equipment:

1. O-E-O regenerators,
2. add-drop multiplexers, and
3. terminal multiplexers.

O-E-O regenerators are used to regenerate optical signals that travel long distances. The most straightforward, although not necessarily the most effective way to regenerate is to convert to the electrical domain and then back to the optical domain, as discussed elsewhere in the book.
SONET O-E-O regenerators might differ from protocol-independent O-E-O regenerators by

implementing additional performance-monitoring functions. To understand performance monitoring, we need to know how a SONET frame is built, which comes later in the chapter. (ADMs) Add- drop multiplexers are the most versatile pieces of SONET networking gear, as they can add or drop any amount of SONET traffic, as desired by the network operations. SONET ADMs are used to create SONET transport networks consisting of SONET rings and point-to-point connections.
(TMs) Terminal multiplexers are a specialized class of ADMs used at the edges of SONET networks. They have the capability to multiplex lower bandwidth signals coming from SONET or non-SONET access networks. Terminal muxes are used to aggregate lower-bandwidth traffic into higher-bandwidth SONET pipes for transmission over optical fibers.
An example of SONET transport architecture is shown in Figure 4.3 .

This particular SONET ring consists of fi ve ADMs, which are collecting traffi c from other ADMs and TMs residing outside the ring. In addition to ADMs, the rings contain O-E-O regenerators which are inserted in the network when the distance between two ADMs becomes too large. SONET regenerators are much simpler and cheaper than ADM regenerators, but unlike ADMs, they have no ability to branch out traffi c. In addition to O-E-O regenerators, optical amplifi ers can be used in the ring as well. -->

Click the Quiz link below to test your knowledge of a design team's responsibilities and the documents used in network architecture.
Evaluating Networks - Quiz