YT YouTube2Text

YouTube2Text → Transcripts → Building Block Approach

Building Block Approach — Transcript

by Santelmo · 8,455 words · 1,393 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:11on this video lecture we will be talking
0:13about the building block approach model
0:16in network design.
0:22So topic outline on this video lecture
0:25includes the introduction about the
0:27building block approach in network
0:29design.
0:30We'll be also talking about the
0:32traditional network design. Okay. So the
0:34building block approach which covers
0:37three phases. These are needs analysis
0:40which covers geographic scope,
0:43application systems, network users,
0:46categorizing network needs and the
0:48deliverables of the pics.
0:51On the second phase, we will be talking
0:53about the technology design. This
0:56includes designing clients and servers,
0:59designing circuits and devices,
1:02network design tools and the
1:04deliverables of the phase.
1:07And the last one is
1:09we will be talking about cost assessment
1:12which includes request for proposal,
1:15selling the proposal to the management
1:17and the deliverables.
1:19All right. So let's get started.
1:23So the objectives of this video lecture
1:25is for you to be familiar with the
1:27process of designing and implementing
1:29the network. So be familiar with
1:32techniques for developing a logical
1:34network design and developing a physical
1:37network design and be familiar of course
1:40with the network design principles.
1:43All right. So let's talk about the
1:46introduction. Okay. So all but the
1:49smallest organizations have networks
1:52which means that most network design
1:54projects are the design of upgrades or
1:57extensions to existing networks rather
2:00than construction of the entirely new
2:03networks.
2:05Even the network for an entire new
2:08building is likely to be integrated with
2:10organizations existing backbone or
2:13nonetheless the network design is very
2:16challenging.
2:20Okay. So the traditional network design
2:24follows a very structured systems
2:25analysis and design process similar to
2:28that used in building the applications.
2:32Okay. First the network design meets
2:35with a user to identify the needs and
2:38application systems planned for the
2:40network.
2:42Second, the analyst develops a precise
2:45estimate of the amount of data that each
2:48user will send and receive.
2:51Uses this estimate the total amount of
2:53traffic on each part of the network.
2:57Next is the analysts
3:00design circuits needed to support this
3:03traffic plus a modest increase in
3:05traffic are designed and cost estimates
3:08are obtained from vendors.
3:11Okay. And finally
3:14a year or two years later the network is
3:17built and implemented. [clears throat]
3:19So this is the traditional network
3:22design. Okay. So this traditional
3:26process although expensive and
3:28timeconuming works well for static and
3:32slow evolving networks.
3:35So unfortunately networking today is
3:38significantly different from what it was
3:42when the traditional process was
3:43developed. So three forces are making
3:47the traditional process less appropriate
3:50for many of today's network. So first
3:55the underlying technology for the client
3:56and server computers networking devices
4:00and the circuits themselves is changing
4:02very rapidly.
4:04So in the early 1990s, mainframes
4:07dominates the network. The typical
4:10client computer was an 8 MHz 386 with 1
4:14MGB of RAM and 40 MB of hard disk space.
4:19And a typical circuit was 9,600
4:22bps or bit per second mainframe
4:25connection or a 1 Mbps LAN.
4:29So today client computers and servers
4:33are significantly more powerful and
4:35circuit speeds of 100 Mbps and 1 Gbit
4:39per second are common. We also have this
4:4210G or even 100G or 100 Gbit per second.
4:48We now have more processing capacity and
4:51the network capacity than ever before.
4:53So both are no longer scars commodities
4:58that we need to manage carefully. Okay.
5:01So second
5:04you've got network traffic is growing
5:07rapidly.
5:08Okay. So the growth of traffic is
5:10immense. The challenge is not in
5:13estimating today's users demand but in
5:15estimating its grow uh rate of growth.
5:20In the past, network demand essentially
5:23was driven by predictable business
5:25systems such as order processing. So
5:28today, much network demand is driven by
5:31less predictable user behavior such as
5:34email and the web. Many experts expect
5:38the rapid increase in network demand to
5:41continue especially as a video, voice
5:43and multimedia applications becomes a
5:46common place on network. So at 10%
5:49growth rate, user demand on a given
5:52network will increase by one/ird in 3
5:55years. So at 20% it will increase by
5:58about 75% in 3 years. So at 30% it will
6:03double in less than 3 years. So a minor
6:06mistake in estimating the growth
6:09or the rate can lead to major problems.
6:12With such rapid growth, it is no longer
6:15possible to accurately predict network
6:17needs for most networks. So in the past,
6:21it was not okay. And finally,
6:26the balance of cost have changed
6:28dramatically over the past 10 years. In
6:30the early 1990s,
6:32the most expensive item in a network was
6:35the hardware or circuits, devices, and
6:38servers. Okay? So today the most
6:41expensive part of the network is the
6:43staff members who design, operate and
6:46maintain it.
6:48So as the costs have shifted the
6:51emphasis in the network design is no
6:52longer on minimizing hardware costs
6:55although it is important. So the
6:58emphasis today is on designing networks
7:01to reduce the staff time needed to
7:04operate them. Okay. So the traditional
7:08process minimizes the equipment cost by
7:10tailoring the equipment to a careful
7:12assessment of needs but often results in
7:16a mismatch
7:18of different devices with different cap
7:21capabilities and capacities.
7:23So two resulting problems are the staff
7:26members need to learn to operate and
7:28maintain different devices and that it
7:31often takes longer to perform network
7:34management activities because each
7:36device may use slightly different
7:39software. Okay. So for instance in
7:42reality today in the network so we have
7:45the hybrid of different brands and
7:47therefore you have to train your staff.
7:49Okay. So to operate to configure and to
7:52maintain these devices for example your
7:54company is using Cisco networks on the
7:56core okay and maybe Junifer or HP okay
8:03on the distribution
8:05and something like uh Dell okay maybe on
8:09the access level so today the cost of
8:13the staff is one more expensive and it's
8:16far expensive than the cost of
8:17equipment. Thus the traditional process
8:20can lead to a false economy. Save money
8:23now in equipment. Okay. So, but pay much
8:26more over the long term and staff costs.
8:31Okay.
8:38All right. So, let's talk about the
8:40building block approach now. So, some
8:43organizations still use the traditional
8:44process of network design. So
8:46particularly to those applications for
8:49which hardware or network circuits are
8:51usually expensive example ones that
8:54cover long distances through many
8:56different countries or regions within
8:58the country. However,
9:01many other organizations now use a
9:03simpler approach to network design that
9:05will be called building block approach
9:07or the building block process. So the
9:09key concept in the building block
9:11process is that networks that use few
9:15standard components throughout the
9:17network are cheaper in the long run than
9:20networks that use a variety of different
9:22components on a different parts of the
9:24network. So rather than attempting to
9:27accurately predict users traffic on the
9:29network and build networks to meet those
9:32demands, the building block process
9:34instead starts with few standard
9:36components and uses them over and over
9:39again. So even if they provide more
9:42capacity that is needed, the goal is
9:45simplicity of the design.
9:48Okay. So this strategy is sometimes
9:51called narrow and deep because a very
9:54narrow range of technologies and devices
9:56is used over and over again. So very
9:58deeply throughout the organization. So
10:01the results are a simple or simpler
10:05design process and a more easily managed
10:08network built with smaller range of
10:11components. So let's get into the
10:13details of the building block network
10:16design. Okay. So in this topic we will
10:20focus on the building block process to
10:22network design.
10:24Okay.
10:27And
10:28the basic network design process
10:30involves three steps that are performed
10:33repeatedly.
10:34Okay. So these are the needs analysis,
10:37technology design and cost assessment.
10:41Okay.
10:49Okay. So let's talk about the network
10:51design here. Okay. So as mentioned
10:53earlier, so there are three phases in
10:55the network design and these are the
10:59needs analysis, technology design and
11:01cost assessment. So this process begins
11:04with needs analysis
11:07during which the designer attempts to
11:09understand the fundamental current and
11:11future network needs of various users.
11:14departments and application. So this is
11:17likely to be educated guess at best.
11:21Okay. So users and applications are
11:23classified as typical or high volume.
11:27Okay. So specific technology needs are
11:30identified. Example the ability to dial
11:33in with the current modern technologies.
11:36Okay. So that's part of the needs
11:38analysis.
11:39Now on the technology design. Okay.
11:42Okay, let's talk about technology design
11:43here. So, this is the next step. This
11:47examines the available technologies and
11:49assesses which options will meet users
11:53needs.
11:55The designer makes some estimates about
11:58the network needs for every category of
12:01users and circuits in terms of the
12:03current technology. Okay. So because the
12:07basic network design is general, it can
12:10easily be changed as needs and
12:12technologies change. So the difficulty
12:16of course lies in predicting user
12:18demands. So one can define technologies
12:21needed. So most organization solved this
12:24by building more capacity than they
12:27expect to need and by designing networks
12:31that can easily grow and then closely
12:34monitoring growth so that they expand
12:36the network ahead of the growth pattern.
12:39Okay. And the last one is cost
12:42assessment. Okay. So this is the third
12:44step cost assessment. The relative costs
12:47of the technologies are considered.
12:50Okay. And in reality the team who is
12:54responsible for needs analysis is a
12:56different team who makes the technology
12:59design from the other team who made cost
13:03assessment. So we have different teams
13:05working on these spaces. Okay. So being
13:08an IT professional. Okay. So for
13:11instance you are part of the technology
13:12design team. So you will be working only
13:15under technology design. All right.
13:23Okay. So the process then cycles back to
13:27the needs analysis. So for instance from
13:29the last page presented earlier. So you
13:31are already on the cost assessment.
13:33Okay. So if there would be some
13:35considerations again. So that would be
13:38cycle back to the needs analysis which
13:41is refined using the technology and cost
13:45information to produce new assessment of
13:48users needs. So this in turn triggers
13:51changes in the technology design and
13:53cost assessment and so on. So by cycling
13:56this three processes the final network
13:59design is settled as shown here. So for
14:02instance let's have an example in here
14:05under needs analysis basically for
14:07instance you need to construct or to
14:11upgrade the network or the current
14:13network infrastructure of your
14:14organization. Okay. And then the
14:17technology design team proposes
14:20specific brands for instance Cisco.
14:22Okay. And then after that after
14:25preparing all the specifications on the
14:28design on the technology design it was
14:30submitted to the cost assessment team
14:32and find out or found out that the cost
14:35is too high. Okay. So it would cost too
14:38much that is why it was forwarded again
14:42to the needs analysis team to do some
14:44adjustments and refinement. Okay. So the
14:47needs analysis team will forward it back
14:50to the technology design and
14:52consider the comment of the cost
14:54assessment team that the cost is higher.
14:59Okay. So therefore
15:01it will be reconsidered by technology
15:03design by proposing maybe
15:06a not too technical or it's not too high
15:11tech. Okay. So devices. So maybe we have
15:14to consider other brands. Okay. So with
15:16cheaper costs okay and then it will be
15:19forwarded again to the cost assessment
15:21team but then if there is a problem in
15:24case for instance for the budget okay so
15:26we have the budget constraint it will be
15:28forwarded back again to the needs
15:29analysis team and so on. So if there is
15:32an agreement now okay so we are okay
15:35with this we are good with this we are
15:37within the budget according to the cost
15:39assessment team then we have come up
15:42with a final network design okay so it
15:44will be forwarded to the technology
15:46design team to come up with a final
15:49network design based on the budget
15:53approved by the cost assessment team
15:55okay so that's how it works
15:58okay so let's dig in into detail about
16:01these three phases. So let's start with
16:03the needs analysis.
16:05Okay.
16:08So analyze the needs of the network
16:10users along with the requirements of the
16:12network applications. Just like starting
16:15with the software development, we need
16:18to start with needs analysis. So the
16:20goal of needs analysis is to understand
16:22why the network is being built and what
16:26users and applications it will support.
16:29So in many cases the network is being
16:32designed to improve poor performance or
16:35enable new applications to be used. In
16:38other cases the network is upgraded to
16:41replace unreliable or aging equipment or
16:45to standardize equipment so that only
16:49one type of equipment, one protocol or
16:52one vendor's equipment is used
16:54everywhere in the network.
16:56Okay.
17:00Next, often the goal of network design
17:03are slightly different between LANs and
17:06backbones or BNS on the hand or on one
17:10hand and mans and ones on the other. In
17:14LAN and backbone network environment,
17:17the organization owns and operates the
17:19equipment and the circuits. So once they
17:22are paid for there are or there are no
17:26traditional or additional charges for
17:28usage. However, if major changes must be
17:33made the organization will need to spend
17:35additional funds in this case most
17:38network designers tend to air on the
17:42side of the building. Okay. So too big a
17:46network.
17:47Okay. So that is building more capacity
17:51than they expect to need.
17:54Okay.
17:57Next is
17:59so at the man one level circuits are
18:01list and upgrades involve determining if
18:04capacity increases are needed. So
18:06basically this man and one is beyond the
18:10control of the organization. Okay. So it
18:13is uh being controlled by the ISP or the
18:16internet service provider.
18:18So in contrast in most man and ones the
18:22organization lease the circuits from the
18:24common carrier and pays for them on a
18:27monthly or per use basis. So
18:30understanding capacity becomes more
18:32important in this situation.
18:35So because uh additional capacity comes
18:39at a noticeable cost. Okay. So in this
18:43case, most network designers tends to
18:46error on side of the building. That is
18:49too much. All right. So there's a
18:53problem with that one. Okay. But it is
18:56much more difficult to cancel a
18:58long-term contract for the capacity that
19:00they are not using. Okay. So much of the
19:04needs analysis may already have been
19:06done before most network
19:10design projects today are network
19:13upgrades rather than the design of the
19:15entirely new networks.
19:17Okay.
19:20So
19:22it is important to gain an understanding
19:25of the current operations application
19:28systems and messages. So this step
19:30provides a baseline against which future
19:34design requirements can be gauged.
19:38It it should be or it should provide a
19:40clear picture of the present network if
19:42one exists. Okay. So existing costs and
19:47user management needs whether the
19:50network is a new network or a network
19:53upgrade. So the primary objective of
19:55this stage is to define first geographic
19:58scope of the network and the users and
20:03applications
20:04that will use it. Okay. So basically the
20:08goal of the needs analysis is to produce
20:11a logical network design. Okay. And well
20:17when you say logical network design this
20:20is a statement of the network elements
20:22needed to meet the needs of the
20:24organization. So the logical design does
20:26not specify technologies or products to
20:30be used. Okay. So take note of that on
20:33the needs analysis we do not specify the
20:36brands. There is no branding also in the
20:39technology design. So when we prepare
20:42the needs analysis and technology design
20:45there should be no brand that is to be
20:48mentioned. All right. So do not specify
20:51brands. You specify only the general
20:55specifications of the device. All right.
20:58So instead it focuses on the fundamental
21:01functionality needed such as highspeed
21:03access network or access to the network
21:06which the technology design stage will
21:08be translated into specific
21:10technologies. All right.
21:13Now before that let's move on to the
21:16geographic scope. Okay. So the first
21:20step in needs analysis is to break the
21:22network into three conceptual parts on
21:25the basis of their geographic and
21:26logical scope. Okay. So these are the
21:30access layer, the distribution layer and
21:33the core layer. So as first discussed in
21:35the next topic. So the access layer is
21:39the technology that is closest to the
21:41user. This is the work area on the EIA
21:45TIA wiring standard. Okay. So access
21:48layer is within the work area. Okay.
21:53Next. So how about the distribution
21:57layer? The distribution layer is
21:59basically the telecommunication closets
22:02wherein all the switches used for
22:05distribution. Okay. So would it be layer
22:07two or layer three switches is placed.
22:11So we call it the distribution layer.
22:12the telecommunication closets and the
22:15core layer which includes the equipment
22:18room and sometimes your backbone is also
22:22considered to be a core layer. Okay. So
22:25which interconnects the different
22:27distribution layer.
22:29All right. So within each part of the
22:32network the network designer must
22:34identify some technical constraint. For
22:36example, if the access layer
22:40is a man in that the users need to
22:44connect to the network over the
22:46broadband connection, this provides some
22:48constraints on the technologies to be
22:50used. So, one could not use 100 bas
22:54Ethernet for example. Likewise, if the
22:57access layer is a LAN, it would be silly
23:00to consider using T1 circuits. All
23:02right? So of course if it is an access
23:04layer we are using the UTP cable or the
23:08100 based Ethernet. Okay. If it is on a
23:11man well basically the end users are
23:13connected directly to the internet.
23:15Okay. So sometimes the current network
23:18infrastructure also imposes constraints.
23:21For example, if we are adding a new
23:24building to an existing office complex
23:26that use 100 bas in the access layer
23:30lands, then we will probably choose to
23:34use uh 100 bas for the access layer in
23:37the new building as such constraints are
23:40noted. Okay. So, it is easiest to start
23:45with the highest level. So most
23:47designers begin by drawing a network
23:50diagram for any ones with international
23:53or countrywide locations that must be
23:56connected. So a diagram that shows the
23:59logical network design going between the
24:02locations is sufficient. So details such
24:05as the type of circuits
24:07okay connected to the one that's also
24:10drawn usually in a series of separate
24:13diagrams but for simple network one
24:16diagram may be sufficient.
24:19Okay. So at this point the designer
24:22gathers general information and
24:24characteristics of the environment in
24:27which the network must operate. So for
24:29example, they determine whether they are
24:33any legal requirements such as local,
24:36state, provincial, federal or
24:38international laws or municipality or
24:40city regulations or building codes that
24:44might affect the network. Okay. So
24:47that's geographic scope.
24:52Okay. So for instance referring to this
24:56diagram here. Okay. So in this case we
24:59have four different branches for the
25:01organization. So one is in Toronto,
25:03Chicago, New York and Atlantic Canada.
25:07Okay. Now this diagram here shows the
25:10initial drawing of a network design for
25:13an organization with offices in these
25:15four areas connected to the core network
25:18which is a wide area network or this is
25:21provided and controlled by the ISO
25:23internet service provider. So the
25:26Toronto location for example, okay,
25:30has distribution layer. So you've got a
25:33BN or a backbone network connecting
25:36three distinct access layers lands. So
25:39which could be three distinct lands in
25:42the same office building.
25:45So Chicago also has a similar structure
25:50with the addition of the fourth access
25:52part that connects to the internet. All
25:55right.
25:57And the Atlantic Canada network design.
26:01Okay. So has two distinct access layer
26:04parts. Okay. One is a LAN and the one is
26:09an access layer man using dialab.
26:13Okay. So the network or the New York
26:17section. So this one here is more
26:20complex having its own core network
26:22component. a backbone network connecting
26:27into the core one which in turn supports
26:29three distribution layer BNS. So each of
26:32these supports several access layer LS.
26:36Okay. So in network uh design or in
26:39needs analysis. So a simple drawing like
26:42this
26:44can use to describe the geographic scope
26:47of the project. Okay. Okay. So start
26:49with writing for instance a branches
26:52where is the main office and then try to
26:54connect them. Okay. So that's the
26:56objective of the geographic scope. Okay.
26:59So connecting your networks from
27:01different locations.
27:04Okay. Now next is application systems.
27:08Okay. So once the basic geographic scope
27:12is identified the designers must review
27:14the list of applications
27:16that will use the network and identify
27:20the location of it. So this should uh
27:23this information should be added to the
27:26emerging network diagrams. So this
27:28process is called baselining. So next
27:32those applications that are expected to
27:35use the network in the futures are
27:38added. Okay. So in many cases, okay,
27:43so in many cases the applications will
27:45be relatively well defined specific
27:48internal applications for example
27:49payroll or registar systems for the
27:52university and external applications
27:56like web servers may already be part of
27:59the old network. So however it is
28:01important to review the organization's
28:03long range and short-range plans
28:05concerning changes in the company goals.
28:08So strategic plans, development plans
28:11for new products or services,
28:13projections of sales, research and
28:15development projects, major capital
28:18expenditures.
28:20What else? uh possible changes in
28:22product mix,
28:24new offices that must be served by the
28:28communications network, security issues,
28:31okay, and future commitments to
28:33technologies. So for example, a major
28:36expansion in the number of offices or a
28:38major electronic commerce initiative
28:41will have a significant impact on the
28:43network requirements.
28:46Okay. So it is also helpful to identify
28:49the hardware and software.
28:52Okay. So
28:55this includes
28:57the software requirements, the hardware
28:59requirements for its applications that
29:01will use the network and if possible the
29:04protocol each application uses example
29:07HTTP. Okay. So for the web access or
29:11HTTP over TCP IP, okay, so the Windows
29:15file access and so on. So all of this
29:17should be considered. So this knowledge
29:19helps now and will be particularly
29:22useful later when designers develop a
29:25technological solutions.
29:27All right.
29:30Okay. So let's talk about network users
29:33now. Who will be using the network you
29:35are designing? Okay. So in the past
29:39application systems accounted for the
29:41majority of the network traffic. So
29:43today
29:45much network traffic is produced by the
29:48discretionary use of the internet. Okay.
29:51So applications such as email and the
29:54web are generating significant traffic.
29:58So the network manager is no longer in
30:00total control of the network traffic
30:01generated on his or her networks. So
30:05this is likely to continue in the future
30:08as network hungry applications such as
30:10desktop video conferencing okay so
30:14become more common. So therefore in
30:17addition to understanding the
30:19applications you must also assess the
30:21number and the type of users that will
30:23generate and receive network traffic and
30:26identify their location on the emerging
30:29network diagram.
30:32Okay. So this is very important. No. So
30:35we have to know the users. Okay. So who
30:39will be the end users of the systems
30:41that we are doing. Okay. So that's part
30:45of the needs analysis. Okay.
30:49And also we have to consider network
30:51upgrades. Okay. So that will require
30:53understanding how the use of the new
30:56application such as video will affect
30:58the network traffic.
31:00Okay.
31:03Next is categorizing network needs.
31:06Okay. So at this point the network has
31:09been designed in terms of geographic
31:12scope, the application systems and the
31:14users. Okay. So the next step is to
31:17assess the relative amount of traffic
31:20generated in each part of the network.
31:22So with a traditional design approach,
31:25this involves considerable detailed
31:27analysis. With a building block
31:29approach, the goal is to provide some
31:32rough assessment of the relative
31:34magnitude of the network needs. Okay. So
31:38each application system is assessed in
31:42general terms to determine the amount of
31:44network traffic it can be expected to
31:47generate today and in the future. So
31:50compared with other applications
31:53likewise each user is categorized as
31:57either typical user or a hightraic
32:00users. Okay. So we have to classify the
32:04users as typical or hightraic users. So
32:09this assessment will be refined in the
32:11next stage of the design process. Okay.
32:15So this assessment can be problematic
32:18but the goal is some relative
32:21understanding of the network needs.
32:23Okay. So some simple rule of thumb can
32:26help. For example, applications that
32:29require large amounts of multimedia data
32:32for those that load executables over the
32:35network are likely to be hightraic
32:36applications.
32:38Okay. So applications that are
32:41timesensitive or need sometimes
32:44constant update okay example financial
32:48information systems order processing
32:51enrollment system okay or the for
32:54instance that the LMS used by the uh the
32:57university okay are likely to be hightra
32:59applications
33:02okay so once the network requirements
33:05have been identified they also should be
33:08organized into mandatory requirements,
33:10desirable requirements and a wish list
33:13requirements. So this information
33:15enables the development of a minimum
33:18level of mandatory requirements and a
33:21negotiable list of desirable
33:23requirements that are dependent on cost
33:26and availability. So for example,
33:29desktop video conferencing may be a wish
33:32list item, right? So but it will be
33:34omitted if it increases the cost of the
33:37network beyond what is desired. Okay. So
33:41you have to have some add-ons you know.
33:43So and then if this will not or if this
33:47will hamper or hinder the design then we
33:51can easily remove it from the list.
33:53Okay. So at this point the local
33:56facility network diagrams are prepared.
33:59Okay.
34:00So for a really large network there may
34:03be several levels. For example the
34:05designer of the network in Chicago,
34:09Atlantic Canada and New York. Okay. So
34:12we have we might have different
34:13designers there. Okay. So conversely the
34:16designer might just add more detail okay
34:21and develop a separate more detailed
34:23diagrams for New York. So the choice is
34:26up to the designer. So provided that the
34:29diagrams and supporting text clearly
34:32explain the network needs.
34:35Okay.
34:37So next is now now that we are done with
34:41needs analysis. So we have to know the
34:43deliverables. Okay. So what would be the
34:45output of this space?
34:47Okay. So the key deliverable for the
34:50needs assessment stage is a set of
34:52logical network diagrams showing the
34:55applications,
34:56circuits, clients and servers in the
35:00proposed network. So it's categorized as
35:03either typical or high traffic as we
35:06mentioned earlier.
35:08Okay. So the logical diagram is the
35:11conceptual plan of the network and does
35:13not consider the specific physical
35:15elements. example routers, switches,
35:18circuits that will be used to implement
35:20the network.
35:22Okay. So
35:24we have here a sample deliverable. Okay.
35:27So sample needs assessment. This is Alan
35:31or local area network. Okay. So it shows
35:35the result of the needs analysis for one
35:37of the New York part of the network
35:41presented earlier. So the figure shows
35:43the distribution and access parts in the
35:45building with a series of six access
35:48lands connected by the distribution or
35:50the backbone network which in turn
35:52connected the campus area network for
35:55BN. Okay. So one of the six lands is
35:58highlighted as a hightra land whereas
36:02the others are typical. So three
36:05mandatory applications are identified
36:07that will be used by all network. Okay.
36:10So these are
36:13email,
36:14web and file sharing. So one wish list
36:18in uh requirement is the desktop video
36:21conferencing
36:22which is also identified for a portion
36:24of the network. So if you will observe
36:27here this is the deliverable for the
36:29needs analysis. So after you have
36:31performed the Toro needs analysis. So
36:33you have to come up with this layout. So
36:36this is just basically a part of the
36:38network design and this focus on the
36:42building design. Okay. So you've got the
36:44campus core backbone there. So if we are
36:47going to relate this into an EIA TIA
36:49wiring standards that we have discussed
36:51on video one. Okay. So this campus core
36:55backbone here is your vertical cabling.
36:58And we also have their cable running on
37:01each of these uh area here the east and
37:04the west land on every layer on every
37:07floor of the building. So that is via
37:09horizontal cable. Okay. So we can relate
37:13this on the first video by adopting the
37:16EIA or the TIA EIA wiring standards and
37:19that should be observed also on the
37:23building block approach.
37:26All right.
37:28Okay. So now that we are done with needs
37:30analysis, let's talk about technology
37:33design. Okay. So this is the second
37:35phase of the building block approach.
37:41Okay. So once the needs have been
37:43defined in the logical network design,
37:45the next step is to develop a physical
37:47network design or a set of possible
37:49designs. Okay. So the physical network
37:52design starts with the client and server
37:55computers needed to support the users
37:58and the applications. So if the network
38:01is a new network, new computers will be
38:04needed. Okay. If the network is an
38:07existing network, the servers may need
38:09to be upgraded to the newest technology.
38:13Once these are designed, then the
38:15circuits and the devices connecting them
38:17are designed.
38:19Okay. So next is how about designing
38:23clients and servers.
38:26Okay.
38:27Now the idea behind the building block
38:29approach is to specify the needs in
38:32terms of some standard units. Okay. So
38:36typical users are allocated the base
38:39level client computers as
38:43server. Okay. Supporting typical
38:46applications. So users and servers for
38:49applications needing a more powerful
38:51computers are assigned some advanced
38:54computers okay or the specifications. So
38:57as the specification for computers
38:59rapidly improve and cost drops usually
39:03every six months. Okay. So today's
39:05typical user may receive a type of
39:08computer originally intended for the
39:10advanced user. Okay. So when the network
39:12is actually implemented and the advanced
39:15users may end up with a computer not
39:18available when the network was designed.
39:22Okay.
39:26Next designing circuits and devices.
39:30So what do we have here? So the same is
39:33true with network circuits and devices.
39:35So habs, routers, switches, okay,
39:38firewalls, these are interrelated
39:42decisions in designing network circuits
39:44and devices. So the fundamental
39:46technology and protocols. Okay. So the
39:49capacity of each circuit these are
39:51interrelated because each technology
39:53offers different circuit capacities. So
39:57when you say circuits and devices we're
39:59talking about the transmission medium or
40:01the transmission media that is to be
40:03used on your network design. So this
40:05devices includes
40:08uh end user devices. These are
40:10computers, printers, okay uh vi phones,
40:14okay and intermediary devices. These are
40:17the network devices like routers,
40:19switches, uh firewalls, bridges and so
40:22on. So this is under this technology
40:26design base. Okay. Now designing the
40:29circuit capacity means capacity planning
40:32estimating the size and type of standard
40:35and advanced network circuits for each
40:38type of network. Okay. So for example,
40:41should the standard LAN circuit be
40:43shared or switched?
40:45Okay. say 100 bas. Likewise, should a
40:49standard backbone network circuit be 100
40:52bas or 1 g uh Gbps or GBE or the gigabit
40:57Ethernet? Okay. So,
41:02this requirements of the current and
41:04future circuit loading the amount of
41:07data transmitted on a circuit. Okay. So
41:10this analysis can focus on either the
41:13average circuit traffic or the peak
41:16circuit traffic. All right. So for
41:19example, okay.
41:22So for example, this analysis,
41:26okay, in an online banking network,
41:29okay, so traffic volume picks usually
41:31are in the midm morning, okay, bank
41:34opening and just prior to closing. So
41:37airline and rental car reservations
41:39network designers look for peak volumes
41:43before and during the holidays or other
41:46vacation periods whereas telephone
41:47companies normally have their highest
41:49pick volumes on Mother's Day. All right.
41:52So designing for big circuit traffic is
41:55ideal.
41:57All right. So Valentine's Day. All
42:00right. So or the Christmas Eve. So
42:03that's the times for the
42:04telecommunications that they experience
42:06heavy traffic because of the greetings,
42:08right?
42:10Okay. So how about the estimating
42:14circuit traffic?
42:17Okay. So the designer usually starts
42:20with the total characters transmitted
42:22per day on each circuit or if possible
42:26the maximum number of characters
42:28transmitted per 2C interval if picss
42:32must be met. Okay. So you can calculate
42:35message volumes by counting messages in
42:38a current network and applying some
42:40estimate cross rate. If an existing
42:43network is in place, network monitors or
42:46analyzers may be able to provide the
42:48actual circuit character okay or the
42:52count of it, okay, of the volume
42:55transmitted per minute or per day. So
42:58you don't have to manually do it. All
42:59right? So of course you have to use some
43:01tools available to make your process an
43:04automated one. Okay?
43:08Now a good rule of thumb is that 80% of
43:11the circuit loading information is easy
43:14to gather. So the last 20% needed for a
43:18very precise estimates is extremely
43:20difficult and expensive to find. So
43:22however precision usually is not a major
43:25concern because of their stairst step
43:28nature of communication circuits and the
43:30need to project future needs. For
43:33example, the difference between 100 bas
43:38and 1 GB is quite large. Okay. And
43:44assessing which level is needed for
43:47physical traffic does not require a lot
43:49of precision. So forecasts are
43:51inherently less precise than
43:54understanding the current network
43:56traffic. Okay. So the turnpike effect is
44:00an expression that means that the
44:02traffic increases much faster than
44:04originally forecasted. So it comes from
44:08the traffic forecasting that was done by
44:11or that was done for the construction of
44:13the early uh interstate highways. Okay.
44:16So when a new faster highway or network
44:19is built, so people are more likely to
44:22use it than the old slow. Okay. the slow
44:26or the old one because it is available.
44:30It is very efficient and provides new
44:32capacities and capabilities.
44:34Okay.
44:36So the annual growth factor for a
44:38network use may vary from 5 to 50% and
44:42in some cases may exceed 100% for high
44:45growth organization.
44:48So although no organization wants to
44:50overbuild the network okay and pay uh
44:55more capacity than it needs in most
44:58cases upgrading the network costs 50 to
45:0080% okay is a good choice but being
45:06under capacity can cause significant
45:08problems. So given the rapid growth in
45:10network demand and the difficulty in
45:13accurately predicting most organizations
45:16in intentionally overbuilt okay to their
45:19networks and most end up using this
45:22supposedly unneeded capacity for the
45:25next three years okay so well some are
45:28using or overbuilding the network but
45:32some it's fine Okay.
45:39Next, network design tools. Okay. So, it
45:43can perform number of functions to help
45:45in the technology design process. The
45:48network modeling and design tools can
45:50perform a number of functions to help in
45:52the technology design process. Okay. So
45:56with most tools the first step is to
45:59enter a diagram or a model of an
46:02existing network or proposed network
46:04design. So some modeling tools require
46:06the users to create the network diagram
46:09from the scratch. So that is the user
46:12must enter all the network components by
46:15hand placing its server client computers
46:19and circuits on the diagram and defining
46:22what each is. Okay. So for example uh 10
46:27bas or the frame relay circuit which is
46:29nearly obsolete with a 1 Mbps committed
46:32information rate. Okay.
46:37So other tools can discover the existing
46:39network. Okay. So we can use this uh
46:41protocols called uh the Cisco discovery
46:44protocol or the CDP or the LLDP which is
46:47used for open or non-Cisco uh
46:51equipments. Okay. So that is once
46:54installed on the network they will
46:57explore the network to draw a network
46:59diagram.
47:01In this case the user provides some
47:04starting point and the modeling software
47:07explores the network and automatically
47:09draws the diagram itself. So once the
47:12diagram is complete the user can then
47:14change it to reflect the new network
47:17design. So obviously a tool that can
47:19perform network discovery by itself is
47:22most helpful when the network is being
47:24designed okay or upgrade okay to an
47:29existing network and when the network is
47:31very complex.
47:34So once the network design is complete,
47:37the next step is to add information
47:39about expected network traffic and see
47:42if the network can support the level of
47:44traffic okay that can be generated over
47:48the network. So simulation a
47:50mathematical technique in which the
47:52network comes to life and behaves as it
47:55would under real conditions is used to
47:58model the behavior of the communication
48:00network. applications and users generate
48:02and responds to the messages while the
48:05simulator uh track the number of packets
48:08in the network and delays encountered at
48:10each point on the network. Okay. So we
48:14can use this tools okay for the testing
48:19of the network design. So some tools
48:21being generated and used okay are packet
48:25tracer, boson,
48:27you also have um GNS3 and a lot of tools
48:32available online. Okay.
48:35Now once the simulation is complete the
48:38user can examine the results to see the
48:41estimated response times throughput. It
48:44is important to note that this network
48:47design tools provide only estimates
48:49which may vary from the actual results.
48:52Okay. So at this bottlenecks or at this
48:54point the user can change the network
48:57design in an attempt to eliminate this
48:59traffic or bottlenecks and rerun the
49:01simulation.
49:02So good modeling tools
49:05not only to produce simulation results
49:07but also highlight potential trouble
49:09spots. examples uh server circuits or
49:12devices that experience long response
49:14times. The very best tools offer
49:17suggestions on how to overcome problems
49:20that the simulation identified. Example,
49:22network segmentation.
49:23Okay, increasing from T1 to T3,
49:26increasing from the current capacity to
49:30a new or better capacity. So there are
49:34lots of tools available that you can use
49:37for the simulations for the testing okay
49:40of your network design.
49:43Okay. Now we come to an end of this
49:47phase two technology design and what are
49:49the deliverables?
49:51Okay. So what do we need to produce
49:53under technology design? So the key
49:56deliverable is a set of one or more
49:58physical network designs. So most
50:01designers like to prepare several
50:04physical designs so that they can trade
50:07or trade off technical benefits example
50:09performance against cost.
50:12So in most cases the critical part of
50:16the design of the network circuits and
50:18devices. Okay. So in some case or in
50:23case uh we have a new network designed
50:25from the scratch for instance. No. So it
50:28is also important to define the client
50:30computers with care because this will
50:33form a large portion of the total cost
50:36of the network. Usually however the
50:39network will replace an existing network
50:42and only a few of the client computers
50:44in the existing network will be
50:46upgraded. Okay. So on the needs analysis
50:51our output is a logical design. Now in
50:55the technology design our output is a
50:58physical design.
51:00All right.
51:03So here's an example of the technology
51:06design deliverable. Okay. So earlier on
51:10we have just the implementation of the
51:13EIA TIA wiring standards on the
51:15building. Okay. So this time we have now
51:19the the specific connections. Okay. of
51:25the work area. So these are the work
51:27area here and we have the
51:29telecommunication closes for each of the
51:31work area maybe.
51:33Okay. So that's based on the design
51:34here. And this telecommunication closes
51:37or switches here are directly connected
51:40to the core on the lower um ground of
51:43the building connected to the router
51:45going to the campus core backbone.
51:48Okay. So we have here a detailed
51:51physical network design as a deliverable
51:54of technology design. Okay.
52:00Okay. Now that we're done with
52:04needs analysis, technology design, the
52:07last phase of the building block
52:08approach is the cost assessment. Okay.
52:12So what do you have in cost assessment?
52:14So the purpose of this step is to assess
52:17the cost of various physical network
52:19design alternatives produced in the
52:21previous step. The main items are the
52:24cost of the software, hardware and
52:27circuits. These three factors are all
52:30interconnected and must be considered
52:32along with the performance reliability
52:34required. So all factors are
52:36interrelated with regard to cost. Okay.
52:41So estimating the cost of a network is
52:43quite complex because many factors are
52:47not immediately obvious. Okay. So some
52:50of the cost must be considered are this.
52:53Okay. So what are this circuit costs
52:55internet networking devices costs and so
52:57on. Okay. So let's start with circuit
53:00costs including costs of circuits
53:03provided by common carriers or the cost
53:06of purchasing and installing your own
53:07cable. Okay. your UTP, your fiber optic
53:11cable, that is part of the circuit
53:12costs.
53:14All right, your connection to the ISP is
53:16also classified to be under circuit
53:19costs.
53:21Okay, so next is inter networking
53:23devices. So this includes switches,
53:26routers, firewalls, okay, bridges that
53:28you're going to use on your network
53:30design. So these are the intermediary
53:32devices. All right, so the next one is
53:36the hardware cost. Hardware costs
53:38includes servers, computers, okay,
53:41workstations, the NIC's or the network
53:43interface cards, the memory printers,
53:46UPS and maybe your backup storage that
53:50is part of the hardware costs. Okay. So
53:54software costs if we are going to base
53:56this on legalities. Okay. And
53:59authenticity of the software then this
54:01is the
54:03said to be the costly part. Okay. This
54:07is the most expensive component under
54:10the network design. Okay. The software
54:13costs for network includes operating
54:15systems, the application software that
54:17will be used in the network and some
54:18middlewares. Okay. You also have the
54:23network management costs. So network
54:25management cost includes special
54:27hardware software and training needed to
54:30develop a network management system for
54:32ongoing redesign, monitoring and
54:35diagnosing of problems.
54:37Okay. So you also have test and
54:40maintenance costs.
54:42Okay. So cost for special monitoring
54:44equipment and software plus the cost of
54:46on-site spare parts. And the last one is
54:50operational costs. Okay. So the cost to
54:53operate the network. Okay. The request
54:56for proposal.
54:58So that is under the operational costs.
55:00So including the salary of the personnel
55:05maintaining and operating your network.
55:07That's under operations costs also. All
55:11right.
55:14Okay. So next is under the cost
55:18assessment is the request for proposal.
55:21Okay. So although some network
55:23components can be purchased off the
55:25shelf. Okay. So most organization
55:29develop a request for proposal making
55:32large network purchases.
55:35Okay. Now RFPs or requests for proposal
55:39specifies what equipment, software and
55:41services are designed or desired and ask
55:44vendors to provide their best prices.
55:47Okay. So some requests for proposal are
55:49very specific about the items
55:53and that are to be provided in what time
55:55frame? In other cases, items are defined
55:59as mandatory, important and desirable or
56:02several scenarios are provided and the
56:05vendor is asked to propose the best
56:07solution. So in few cases RFPs
56:11specifically or generally what is
56:13required and the vendors are asked to to
56:16propose their own network designs. Okay.
56:19So we have here an example content okay
56:23of the request for proposal.
56:26Okay. So basically
56:29once the vendor have submitted their
56:32proposal the organization evaluates them
56:34against the specified criteria and
56:36select the winners. Okay. So I mean the
56:40who won the bidding process. So
56:43depending on the scope and complexity of
56:45the network, it is sometimes necessary
56:48to redesign the network on the basis of
56:51the information in the vendor's
56:53proposal.
56:54Okay. So
56:57one of the key decisions in the request
56:59for proposal process is the scope of the
57:02request for proposal. Will you use one
57:05vendor or several vendors for all the
57:07hardware, software and services? So
57:11multi-endor
57:12okay environments tend to provide better
57:14performance because it is unlikely that
57:16one vendor makes the best hardware
57:19software and services in all categories.
57:22So the multi-endor networks also tend to
57:25be less expensive because it is likely
57:27that one vendor will always have the
57:29cheaper or the cheapest hardware
57:31software and services and all the
57:34product categories.
57:36Okay. So the request for proposal
57:38includes this background information
57:40which includes uh organizational
57:42profile, the overview of the current
57:44network, the overview of the new network
57:46and the goals of the new network. The
57:49network requirements okay is also there.
57:52The choice of set of possible uh
57:54possible network designs
57:57noted as mandatory, desirable and wish
57:59list items. Security and control
58:03requirements, response time requirements
58:05and guidelines for proposing a new
58:07network designs.
58:10The service requirements, the
58:11implementation time plan, training
58:14courses and materials and support
58:16services, reliability and performance
58:20guarantees.
58:21Okay. So also included is of course the
58:25bidding process. Okay. So time schedule
58:28for the bidding process, the ground
58:30rules, bid evaluation criteria and the
58:34availability of the additional
58:36information.
58:38Okay.
58:40And information required from vendors.
58:42So vendor corporate profile experience
58:44with similar networks uh hardware and
58:46software benchmarking reference lists.
58:49So as I mentioned earlier, the multi-
58:51vendor environments can be more
58:53difficult to manage. However, if
58:55equipment is not working properly and it
58:58is provided by two different vendors, so
59:01each can blame uh the other for the
59:03problem. In contrast, a single vendor is
59:05solely responsible for everything.
59:08And in reality, in the network design,
59:10so we do not adopt a single vendor uh
59:13network. Okay? So because we don't want
59:16to we don't want our network to be
59:19monopolized by one vendor or single
59:20vendor. Okay? Okay. So as much as
59:22possible almost all organizations okay
59:25are using an open systems. Okay. So they
59:28are not after proprietary wherein they
59:31have to pay charges okay with the
59:35software hardware and so on including
59:37supports.
59:39Okay. Now, so this is a big challenge
59:43for you becoming
59:46a network designer or maybe a network
59:49administrator wherein you are going to
59:51sell the proposal to the management. So
59:53how will you convince the organization
59:55to support your proposal of building a
59:57new network. Okay. So an important
1:00:01hurdle to clear in the network design is
1:00:03obtaining the support of the senior
1:00:04management. All right. So gaining
1:00:07acceptance from the senior management
1:00:08lies in speaking their language and
1:00:10presenting the design in terms easily
1:00:13understandable issues. Okay. So one tip
1:00:16is that when you are selling your
1:00:18proposal to the management you have to
1:00:20use the layman's term. You have to use
1:00:23the term okay that is used on their
1:00:26level. If if you're presenting it on a
1:00:28business level then you have to use uh
1:00:30terms like how will the organization
1:00:33benefit from it. Okay. So what are the
1:00:35return of investments? What are the
1:00:37impacts of this proposal to the
1:00:39production or to the income of the
1:00:41organization? So that's how you present
1:00:44it to them. Do not present technical
1:00:46matters on this people because they
1:00:48don't want it. All right? So this
1:00:51technical people only wants how can we
1:00:55increase the profit of the organization.
1:00:58So you have to answer that way. Do not
1:01:00give them out. we we we we can improve
1:01:03the income or the profit of the
1:01:06organization by proposing this Ethernet.
1:01:09Okay. So something like uh we will be
1:01:11using the fiber optic technology. What
1:01:14are those? All right. So what I'm saying
1:01:16is when you are selling the proposal to
1:01:19the management you have to use terms
1:01:22that is understandable on their level.
1:01:26Okay. So rather than focusing on
1:01:29technical issues such as upgrading to
1:01:30giga Ethernet, it is better to make a
1:01:33business case by focusing on the
1:01:34organization's needs and goals such as
1:01:37comparing the growth in the network use
1:01:39with a gross in the network budget.
1:01:41Okay. So
1:01:44one of the main problems in network
1:01:46design is obtaining the support of the
1:01:48senior management. Okay. So that's
1:01:49always the challenge. That's what I'm
1:01:50saying. So to management, the network is
1:01:53simply a cost center. Okay. something on
1:01:57which the organization is spending a lot
1:01:59of money with a little apparent change.
1:02:02Now the network keeps on running just as
1:02:04it did the year before. So the key to
1:02:08gaining the acceptance of the senior
1:02:10management lies in speaking the
1:02:12management language is what I'm saying
1:02:14earlier. Okay. So it is pointless to
1:02:17talk about upgrades from 100 100 Mbps to
1:02:201 GB. What's the sense of that? I can't
1:02:23understand that as a boss. Okay. on the
1:02:25network backbone because this
1:02:27terminology is meaningless from the
1:02:29business perspective. Okay. So a more
1:02:32compelling argument is to discuss the
1:02:34growth in the network use. For example,
1:02:38a simple graph that shows the network
1:02:40usage growing at 25% per year compared
1:02:44with the network budget growing at 10%
1:02:47per year.
1:02:49Presents a powerful illustration that
1:02:52the network costs are well managed.
1:02:55and not out of normal. Okay, so that's
1:02:58how we convince our superior. Okay, so
1:03:01likewise a focus on network reliability
1:03:04is an easily understandable issue. For
1:03:07example,
1:03:08so if the network supports a mission
1:03:10critical system such as order processing
1:03:14or moving point of sale data from retail
1:03:17stores to corporate offices, it is clear
1:03:20from a business perspective that the
1:03:22network must be available and performing
1:03:26properly or the organization will lose
1:03:29revenue. All right. So that's how we
1:03:31sell it to the uh to the management.
1:03:34Okay. So we should have the power to
1:03:36persuade. All right. So it's very hard
1:03:39to persuade people especially the
1:03:41bosses. Okay. So that's why you have to
1:03:43speak on their language. You have to
1:03:45speak the business way or the business
1:03:48level. All right. So don't give them two
1:03:50technical matters, two technical uh
1:03:53terms. What is that? Right. Now what are
1:03:57the deliverables under this space?
1:04:01Okay. So there are three key
1:04:03deliverables for this step. So the first
1:04:06is the request for proposal that goes to
1:04:08potential vendors.
1:04:10All right, that's the first one. The
1:04:12second deliverable after the vendor has
1:04:15been selected, it is or is the revised
1:04:18physical network diagram with a
1:04:20technology design complete.
1:04:23So exact products and costs are
1:04:26specified at this point. Okay. So for
1:04:29example, we will be talking about the
1:04:31port density of the switches. When you
1:04:33say port density, how many ports are
1:04:35there? Would it be 16 or 24 ports or 48
1:04:39ports and so on. All right. So the third
1:04:42deliverable is
1:04:45the business case that provides support
1:04:47to the network design expressed in
1:04:50business objectives. All right. So this
1:04:52is very important. So everything you do
1:04:55in a business, it should be in
1:04:57accordance with the business objectives.
1:04:59All right. So that's it.

About this transcript

This page contains the full transcript of Building Block Approach by Santelmo, generated from the public captions YouTube serves with the video. The transcript has 8,455 words across 1,393 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to translate the transcript into 100+ languages.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.