Friday, May 27, 2022

(Free CCNA Class#3/21) Tier 3, Spine leaf Architecture, On-Premises vs Cloud | “21_DAYS_CCNA”

Okay hello yeah hi we'll just wait for another one minute and then i'll start okay yeah sure okay let's start then hello everyone good evening good morning you're doing good yeah okay okay so let's first let me first talk about what we could .

Complete so far okay we did all the networking devices we understood the installation of packet tracer and we did some simulation of client server model and briefly we have understood what is osi and tcp model a seven layer osi model and a five layer tcp protocol .

Suit okay today i will talk about more a little about two tier three tier and spine leaf architecture and something related to lan van man and a bit of ipv4 addressing okay so today the class is going to be a bit short okay since i got some urgent work today before that .

I have some sort of you know just to start about i just want to know how some questions like few answers from like the given question okay let's say i ask you question number three i hope my screen is visible to everyone so you can just let me know through chat box the answer for question three .

Acknowledgement sequency and flow control are characteristics of which layer come on guys quickly tell me if you have attended the lectures seriously then you should be able to tell me the answers yes very good people are saying option number four which is layer seven that's .

Really good that's really good that is transport layer that's nice okay uh that's nice now let's go to the question number let's say seven when the data is encapsulated that's good guys that's really a good response which i'm seeing from everyone uh question number seven let's go to .

Question number seven uh when the data is encapsulated which is the correct order when it is encapsulated okay that means it's going from the top to the bottom from application layer towards the physical layer okay ayushi says option d naresh says option c .

They all look very alike okay guys so just check out properly what you're saying okay c fewer again saying some of some are saying d okay so majority i see people see are we giving me answers option c okay so first is data then comes a transport .

Layer where we see it it is known as segment then comes the network layer which is nothing but packet then comes the frame and then comes the bit okay so very good option number c is the correct answer okay please remember the layered .

Architecture just for the sake like we have application presentation session transport network data and physical layer so we can make some sort of you know and some sort of acronym or some sort of mnemonic out of this to understand which layer we are talking about .

We cannot miss we cannot just shuffle it okay layer one layer two layer three layer four layer five layer six and layer seven okay so option c is my correct answer okay let's go with the option eight question eight if you use either telnet or ftp which is the highest layer you're using .

To transmit the data okay what do you think when you are using telnet or ftp now telnet is a remote application we have not discussed so much but uh when we were talking about the services we were we were trying to talk about this various .

Services which are there okay omkar says layer 7 uh very good option d that's right so layer 7 is the application layer so all these are the application services very good very good so you all are able to tell me at least you have you're attending very sincerely .

One last question we'll take now uh which protocol is used to find the hardware address of a local device now this was the question which was asked by one of the student in the previous class and i answered it very shortly because i was supposed to cover this topic .

Okay the protocol used to find the hardware address okay i i generally know the ip addresses now i want to find the hardware address okay so a b shake is saying arp most of you are saying option a but not majority yeah okay okay so yes it's ar ph acid it is not icmp okay it's a rp .

Protocol okay icmp is a used for checking the connectivity or you to check the network layer status it's the icmp okay it's the arp sorry it's the arb address resolution protocol where we have a table which maps the ip address with the mac address okay yes roman you're right .

Correct so we have such question answer sessions every day before we start okay so that you we also get to know whether you people are understanding whatever i am teaching okay so i i guess everyone is going good so let's get back to it today's agenda what we are going to learn okay so starting with .

The concept of we have this concept of bandwidth now this was also something which i tried discussing in the previous class about bits and bytes so this is one of the very crucial parameter in the network okay we are always requiring a good sufficient bandwidth so that we get a good quality of .

Experience or a good quality of service because these are the two other parameters which generally we are interested to calculate the quality of experience the quality of service there should be no latency all this there should be no delay and the bandwidth should be as high .

Whatever is this the underlying technology whichever we are using so accordingly the bandwidth should be available to us so it's a very very important parameter it's not available for free we are going we are paying for it and it's very critical for the network .

Performances we know how it has been measured in terms of bits per second kilobits per second megabits per second gigabits per second so it's bits per second now as i was trying to discuss about the three-tier or the two-tier architecture in the previous .

Class i just thought let me give you a very quick overview about land van and man so that we have a clear picture because generally we don't understand a basic uh way of you know under defining what is a lan or a van or a man so i just put down the very simple term .

That we if we have one two or two computers in our home or in our room or in our entire building or in our entire campus and we want to connect them so we say that we are connecting them over a local area network okay so there is quick and simple definition two or more computers .

Or any devices because in it it is not necessary we have computers you are in place we may have switches also so any nodes okay two or more nodes a simple term i can say is nodes that are in a room on a floor in a building or in a campus if they are .

Connected they are connected on land okay some of the features will talk about land that who is the owner of the land what is the cost of deploying the land what are the equipments or the components required for the land so these are the characteristics of the land okay .

That we are supposed to get a very high bandwidth high speed high capacity when we are deploying a lan okay and whoever is the owner of that premises for example when you are setting up a lan at home or in your organization in your company wherever .

Owner of the premises is the owner of the land okay so since we are not taking help or we are not purchasing it someone from some service provider okay we are not going to any service provider to come and you know design a land for us okay so the owner is itself the owner .

The owner of the premises itself is the owner of the land and very important thing what we require is that when we deploy the land the cost of deployment should be very cheap the equipment should be very compact and should be powerful enough to provide you high bandwidth high speed and high .

Capacity and the maximum size of the campus can be defined as the size of your land which is your campus wide land so more or less we try to uh say that we are designing a campus network okay most of the time when we say that a campus network has to be designed .

Or an enterprise network has to be designed so what is this okay basically we are trying to design a land only okay we are trying to design a land okay however when we go on connecting two of our branch offices which are not in .

The same way which are not very close which are not at a very close proximity or which are separated at a very uh a little i can say geographically separated but they are in the same city or they are not in the same city then we move on towards a different .

Approach which is your man and man okay so again we'll understand the characteristics of a man and man the definition will almost remain the same with some with a little modification okay so that you all can remember it very properly it's the same thing but now we are saying that they are geographically .

Separated now this is metropol metropolitan area network we don't talk about this much we either go with the lan or a van network so this is like within the same metro city but geographically separated then they are said to be man and van has to be what it has to be geographically separated .

But not within the same city so that is the definition of a van so this is the general definitions of land man man since we studied or we just saw the characteristic of land that it should have just three important things what we could identify from here that we .

Should get very high bandwidth over a land the owner of the premises is the owner of the land and the cost of deploying is cheap these are the three features of a land now on the same line we'll understand what are the features of man and when okay so let's discuss that .

First so here comes the the characteristics of a man and a one now i have clubbed it together since i have to they are not much different geographically within the same city but they are geographically separated okay that is something which is important within the city or with the behind the .

City that is how they that makes a difference so now we cannot say that we'll have a very high bandwidth over a man or a bank network okay it typically depends on what kind of service you're running okay what kind of service you want and based on that you decide what kind of bandwidth needs .

To be you calculate your bandwidth requirement and accordingly you set up the link okay so we can say that we can have high to moderate bandwidth interconnectivity since the cost is one of the very important factor when you go and purchase a link okay when you go and purchase a link from the service .

Provider so the cost factor comes over you so accordingly you try to compromise or you try to see what kind of investment you can make owner is again owner of the premises is the owner of man and van if it is my organization which and i have several branch offices .

Then the the cost of whatever i can have my own equipments okay owner now see it it is completely dependent on me whether i just i can have my own equipments components which i can use and i can just purchase the link okay if the bandwidth what i'm talking about or i can i can have that .

Both everything the link as well as the uh equipments whatever is required to build a network can be provided by somebody else that is your service providers so mostly we say that if i am holding i am building my own man horowan network i am the owner of .

That particular premises where links are released from the service provider the cost of deploying is very high as links are released from service provider and the customer premises equipment can be leased or owned by the owner of the premises okay it's just like when you go and purchase a 2gb link or a 3gb .

Link or a 5gb link you you sign an agreement okay we say it has an sa okay that is the lease agreement we sign so that we say that we are running so and so services for example i'm a bank organization okay i just take an example that we are we are our organization is a bank .

Okay now we will try to understand what services the bank does mostly because more or less we are dealing with three types of services generally we have a data service we have a voice service we have a video service i don't think so any other thing people okay these are the different three .

Services which we have okay so can you people just tell me through chat which of the services will a bank require bank require more cost of sales data all right that's right data services that that's right maximum okay if i if i see even voice is required that's right .

But maximum we would require a data service followed by if i have to organize some conference meeting with my colleagues or with somebody who is in other branch or at my headquarters then i would definitely require to have a voice conferencing or a video conferencing but the amount of those services which i .

Would use would be very less as compared to the data services so based on that i will calculate like for data how much of bandwidth i need for voice how much of bandwidth i would require for back for video how much of bandwidth of .

I would require and then i'll make a total calculation that okay this is something which i figure it out that this is my requirement okay so based on that because every every services okay every services has something like uh for example when i say that i'm using a video service so i know what .

Is the maximum or like size of the file so accordingly the size of the file and the calculation we can do in terms of bandwidth that according to one uh frame or one video file or one voice how how can we calculate a basic the basic unit of the data voice and video is .

Known to us so based on that we can do some calculations i'll show you something interesting related to this maybe in tomorrow's class uh so these are the services which we have so based on that the bank has to calculate and then you know go to the service provider or check .

The service provider that this is what is my requirement and then you know you can get quotations and these are the links what we provide and this is the service what i will give you i will ensure that your links are up for so and so many time because you need the services to be .

Highly available to you highly uh reliable the service whatever or the links which you are purchasing okay the agreement says that we are going to give you so much of some support like 99 times or 99.99 of time all your services will be available there will be no downtime .

So 0.001 whatever they are saying the downtime accordingly we we sign up in agreement and if that particular thing they fail to achieve that objective or that availability whatever they have said then they got to bear the cost okay so yeah that's right sls so you're in in that in those terms only we are .

Talking so that is why i said the cost of deploying the land is very uh man and van is very high and uh it's like totally dependent on your calc on your service whatever you are running according to that so services of your network defines the architecture of network that is what i .

Just said and the next generation manual network it uses a converge architecture it should be capable of providing video voice services over a converged architecture that means we don't lay down a different network for a data service for a voice service or for a video service we .

We are architecting a network so that all the services can utilize the same architecture so that is what is the next generation converged architecture so converge meaning it says itself that you are trying to take up all the three services over one single architecture now the .

Administrative control of a man can be centralized or a distributed the control it means what like whom should you approach when there is a kind of failure or there is some sort of uh the system has got crash or something if anything happens so who is going to .

Control okay it's like the control is given okay either it can be a centralized manner or it can be in a distributed manner now for a van it is obviously it is required that you need a distributed approach you cannot have one centralized approach okay you need a distributed approach .

Just for an example uh if something happens in your bank okay you go to a bank or uh there's some services which is not available so whom do you approach next do you directly go and approach to the headquarters of the bank or you directly try to approach .

Somebody who is who can you know give you the solution for that so accordingly there is a distributed approach which is followed over over a van or uh and in a man it can be distributed or centralized based on the mission criticality that what kind .

Of services or what is running based on that so i hope now you all have understood what is land van maan because this is very important when we are designing a network so accordingly the kind of structure okay now the two tier and the three tier let's come on to that because now since .

We have understood what is lan man and man i'll just discuss about the 2d or three tier architecture okay now that's that's okay so i hope the packet tracer is visible to everyone is it visible guys .

Okay okay that's cool so now what i see here i have i have just drawn the architecture for you all okay that is we have a three-tier architecture on my left side and uh spineless architecture on my right side there is obviously one more thing which we have a two-tier architecture so starting from the bottom line okay of .

A three-tier architecture so we see that i have three uh i have three types of systems okay i have two desktops and then two laptops and again two desktops now these are nothing but we know these are our end devices or our end points okay so i can have any devices i can have a wireless wireless .

Communication i can also have a wired communication now these are the devices which are connected on the first layer of switches which are known as the axis layer switches okay so the switches which we have which i have taken which is a two nine six zero .

Switch so if you can see that all these switches will be considered as a layer two switches and they are known as the axis layer switches okay so what is the job of the access layer switches the job of the access layer switches is to provide connectivity to all the .

End devices so all my end devices are going to come and connect to this layer of switches so what is the requirement now what do you think is the requirement when you have a layer to switch or access layer switch because you just need to connect or you need to .

Serve all the users because all users are going to connect to you so what is one of the important requirement when you want a layer to switch or an access layer switch anybody who can guess what would be one of the important requirement if i have to connect hundreds and thousands of .

Users okay and all those users are connected on this access layer switch so one of the important requirement of this switch is going to be what it should be what anybody who can guess your voice is breaking whoever is saying your voice is breaking .

Okay you can mention through the chat i guess i could not hear you could you please repeat again my question is that since i said that i am connecting all my end devices over a layer to switch or an access layer switch which is generally uh where all end devices are connected .

So what would be one of the requirement of those l2 switches or access layer switches it should have a connection to the layer 3 switches capable of connecting to there that's that's that's that's important and what what else what is something more important .

Because if you have more and more of you yes you should have a good density port density we say you should have good number of ports so accordingly the one of the requirement is that your ports which you have the port density should be sufficient enough to connect all to the .

All the users the bandwidth okay since somebody said that okay now at the one side you are connecting the end devices at the other side you are connecting to the distribution layer now why do you need that switch okay and we say that distribution layer switch which is above the axis layer switch which we say it .

Again as a layer 2 okay now why do we need that switch when we have to communicate let's say pc 0 wants to communicate with pc 2 which is just a side of that do you think the distribution layer switch has any role to play it doesn't have any role to play right right on them it doesn't have any rule .

Because they are in the same they are connected on the same switch so when they want to communicate they can directly talk with each other with the help of access to your switch but let's say i want to connect with pc three or i want to connect with laptop zero or a laptop one .

Then i my access since there is no connection between the axis layer switches this is something very important you have to understand that the first layer okay since it is known as three tier this is known as the layer one okay where my axis layer switches is there okay so here you can see that in this .

Layer switches they are not interconnected with each other okay they are like acting like a separate like you have to consider that you are designing a network or you're designing a lan to connect to buildings or to connect the flows inside your own building .

So accordingly when you want to connect to a laptop or a pc or a user who is sitting somewhere else okay who is sitting on some other switch okay some other access layer switch then you need somebody to communicate so that person with that switch is going to be your distribution layer switches so all .

The distribution layer switches you will see that there is a kind of mesh connection all access layer switches have access to all the distribution layer switches it there is a connection going from this particular pc this switch to this one this to this one so this is a kind of mesh connection .

Which we say that every this access which has a connection to all the other distribution switches so that we can talk to any of the host or any of the user who is sitting somewhere else okay so that is why this switches okay now whichever we are connecting .

To the distribution layer switches now this is going to be your uh the ports which we say okay the ports whichever is connected to the distribution link layer we say they links and they need a good bandwidth okay there has to be a good bandwidth okay to be supported so that your traffic can be taken off .

Properly one thing which you will you will observe here okay i want you to observe this thing that there are some green links and there are some red or orange what you can say those types of connections you can see you okay now those connections are blocked connection whichever you .

See as a uh orange color okay wherever you see an orange color link it's basically a block connection because we'll learn sometime later about spanning tree protocol that whenever we connect switches and if there is a kind of looping which happens okay we'll come to this in detail just to .

Give you one quick overview that whenever we connect switches and if there is a kind of looping which happens in order to make the thing look free we have a spanning tree protocol so as soon as you connect all i have not configured anything in this switches okay i've just .

Placed them and i've just connected them with the help of cables so even you can go and try this okay just pick up the computers the switches 2960 for your access layer and all the multi-layer switches for your distribution and core and just place them and connect them you will find that there .

They will take some time to turn green or to come up okay in in technical way we say to come up okay so that time of some time later you will see that not all the links or not all the ports are going green okay they some turn out to be orange and that .

Is because the spanning tree protocol was running and it identified that there was something loop which was forming so by blocking one of the port it has made the loop free connection okay and we are going to understand that how it makes the loop .

Free connection and what happens if there is a loop okay what is the drawback of having a loop and what is how do we avoid it okay so spanning tree protocol helps you avoid a looping so what happens over here now since you are trying to connect okay you are trying to connect all your .

Access layer switches to all the distribution layer switches okay there is a possibility that you're from layer 2 layer 1 to layer 2 or from access to distribution layer all your links may not be up okay because due to this loops okay you may have a possibility that .

Some of the port may be blocked so one thing what you see is that you cannot get high amount of bandwidth so all you're connecting three four links in this direction okay and if any of loop is being formed so that particular connection will be blocked so you're not getting high bandwidth .

That is something which i want to say okay so you will not get a high bandwidth when you are trying to you know try to get your traffic passed through the distribution layer switches this is required distribution layer so it is aggregating one of the other term which is used is it is aggregating all .

The access layer switches and then it is trying to forward your traffic based on whatever you know uh you want to communicate with the other one so here you need a sort of routing also okay that is why we have a multi-layer switches we cannot just do with the help .

Of layer 2 switch because layer 2 switch the requirement is just to be a little cheaper because you want to connect all the end devices thousands of users you need good ports with good bandwidth which it it should have for connecting towards the user as well as to connect to the .

Distribution measures so based on that you connect any model like 29502960 you can use any of this for access to your switches but when you move towards the distribution layer switches since you want to communicate the to the computers which are on different access layer switches .

So you need that there has to be a routing routing also which will come into picture okay routing will come into picture switching will come into picture because they may be on different networks it is not necessary that this particular pc and this pc are on the same network .

So you would need routing as well as switching so you make use of the layer three switches or i can simply say the catalyst switches three five five zero three five six zero these are the models which are used so again the requirement of distribution layer switches tends to be like .

The features it requires is it requires a good bandwidth it uh it requires a good bandwidth so that it can you know uh forward the packet as soon as it receives from that it can just take the decision of forwarding the packet to the correct uh access layer so it requires good .

Processor all these things now again you will require something which is if you want to communicate between the different distribution switches now i have three distribution switches in my place let me just remove this uh labeling which i was showing there now it seems a bit better okay so now .

When i want to communicate okay now if i have you can see that i have three distribution switches now if i want to commit so one and two are also interconnected two and three are also interconnected now why do you need so many distribution switches you will have to think .

There is a sort of redundancy which we have kept okay when we are designing a network we want that they should be highly available this is one of the very important requirement guys that you need the network to be highly available okay there should not .

Be much of downtime and in order to achieve this objective of highly available you ensure to keep redundancy in your network okay you ensure to keep a redundancy in your network so that if any time one of the switches so if this switches if this distribution layer switch fails still my .

Packet or my traffic can be routed via different distribution layer switches so you see that there is a redundancy in the distribution layer however i have kept one core layer switch now coming back to the core layer because if you have to connect or if the traffic has to .

Be routed between two distribution layer switches then you need a core layer which is the backbone okay this is the backbone of your entire network of your campus okay your two buildings you can assume that there are two buildings you're developing a land for designing a land .

For that so your core layer switches are connecting those two lands your two land segments two or more lands okay so your core is going to just give you high bandwidth okay it's a backbone of the layer it is not supposed to .

Take any kind it is not going to do any kind of manipulation it is just be involved in forwarding the packet at a higher bandwidth okay so the job of the core layer is to give you good band the switching has to be very fast and it also uses routing protocol to route your package .

So majorly you can say that to connect the access layers you need the distribution layers and to connect the distribution layers you need the core layer okay and there is a redundancy at the core layer there is a redundancy at the distribution layer .

When i talk about redundancy you can see that when the traffic from this layer to this switch doesn't go if it fails due to any reason then my disc distribution switch can take care of it because i have a redundancy i have kept a redundancy i have kept a .

Redundant link okay and i have made it in a full mesh connection so my traffic can be taken up properly by another switch okay so redundancy is observed in the distribution layer it is also observed so i can i have to place one more core layer switch so that if this layer if this course .

Which fails just understand that if you just have one core layer switch and if that fails then your entire communication will be halted okay so this is very important that you see a three-tier architecture or a hierarchical architecture .

Because initially when the system came up okay at the when you can say that two or three decades before when before this structure came up okay it people used to just randomly try to connect all the you know the network was not that uh specific the architecture was not .

Very specific it was like okay you take one switch and you connect now all the devices to one switch but now when the requirement goes more then you take another switch then you take some users there some users here connect the two switches what now if more requirements come then .

How to connect the third switch or the fourth switch or the fifth switch so there was some sort of you know there was not a kind of proper organized way of architecting the network so this was the three-tier architecture or the campus network design which we used to follow .

Okay so that we see a proper structure we see that the proper traffic okay there is a clean and neat way of understanding how the packets will get forwarded how it will get routed and it also helps you scale okay when you see that now if i have to add more of my access layer .

Switches or distribution layer switches so it the scalability is also quite good okay now what is the problem over here is that you see that first and foremost the bandwidth requirement okay the bandwidth whichever since you have a redundancy link you see there is a sort of loop which has been formed now .

We are going to come across loop what it what is a loop as of now you just understand that you see whatever wherever you see the orange link okay wherever you see the orange link that means those links are being blocked they are not participating in communication .

At present they are being blocked because there was a loop which was formed and because of that the link is being blocked now since the link has been blocked you don't get the required bandwidth okay when you suppose if you have four links which is going out of that thing .

And if one of so each port is supporting you 10 let's say 10 gbps of speed each port is giving you okay each port if from every uh every switch if there are four links which are coming out of it and each link or each port is of 10 gbps so total i should be getting a 40 gbps of .

Bandwidth to route my traffic okay however if one or two of them gets blocked because of the loop technology or the loop thing which happens so what happens is the amount of bandwidth which i get would be reduced that is so simple to calculate if one of the things if one of the port gets .

Blocked i'll get 30 gbps if two gets blocked i'll get 20 gbps so likewise because of spanning tree what was happening here that the amount of bandwidth was not up to the mark what was required now for that people went on to the other concepts like ether channel .

Which again you're going to learn how to overcome this kind of scenario where you see that the amount of bandwidth utilization is not good due to the spanning tree protocol which has blocked due to make it loop free okay so there are some other ways other technology methods to overcome that but .

At prison you can see that okay i am not able to get all the required bandwidth which i should be getting for my traffic the second thing what you see is that if a packet has to go from pc 0 to an any of the pc here okay let's say here is my pc zero and to here is my pc one so if a packet .

Has to go or a traffic has to go over here it has to first go to the axis then to the distribution layer from the distribution layer since now this is not directly connected okay however if you want you can actually send your traffic if it is not .

Blocked you can actually send it to this distribution layer switch from there it will go to this access layer and then come to this pc so there's a kind of you know a kind of latency will be involved because you are hopping your packet from the access to the distribution from the distribution back .

To the uh now since i'm using this distribution layer so you see there is not much but if i have to go from the distribution then to the core and then from core to distribution from then distribution to access so you can see if i take this way okay from this axis to distribution to .

Core then back to the distribution then to the axis and then coming back so there will be a lot of latency involved in such a scenario so there were two disadvantage there are many more but first is as of now these are the two disadvantages of this three-tier .

Architecture now people went on to the two-tier architecture by just collapsing the core means if i just remove this core okay let me just clear this drawing if i just collapse if i just why we are not going to the switch multi layer switch one to directly to the switch .

Uh pc to one we can go that way but let's say this link gets max i'm i'm i'm guessing the worst case let's say now there is uh the worst case now this link this this particular switch has got the switch has failed or some link has gone down .

Okay okay so the link may go up and down now we are not connecting we are considering the this year as a very simple design but in a real-time scenario you know we know that there are thousands of users thousands of hosts and you are trying to build such kind of hierarchical model .

Okay so okay in that case i'm just trying to see that what could be the latency involved and secondly what i was saying here that if their people also went on to a two-tier architecture now two-tier architecture also known as the core collapsed or a collapsed core architecture .

Whereby i just remove this and i make this itself as my core as well as the distribution layer switch so when should you go for a two-tier or a three-tier if you have a very small campus if you want to design a very small campus network then definitely you will not go towards the core layer .

Just understand one thing guys that when we are talking about access layer distribution layer or core layer switches their switches are very very expensive because they are going to give you good bandwidth they are going to support both routing as well as switching the distribution as well .

As well as the core layer switches so they are not very cheap as what you get the access layer switches so if you feel that your network design is very small then we better go for a two-tier architecture where the collapse standard the core and the distribution they acting like the same .

Okay and then coming back so people used to follow this network design two tier and three tier way a lot okay however now some scenario has changed with because of this fine leaf architecture now the spine leaf architecture is majorly useful if you .

Are going for a data center okay now anybody who knows what is a data center anyone who can who just can let me know what is a data center have you heard about this term data center yeah where all the information is hosted information is hosted in the form of what all the data .

Or physically hardware tv is everything it's well secured will secure premises area where all the devices may be computers servers storage devices right everything right right so it's just housing all your servers computing devices storage devices storage servers .

Computing servers your networking your switches routers so all it can be just one entire building or it can be just one small room okay so that is what we say that it is just a housing of all your servers and networking components okay so all the facility .

Is being arranged in such a manner that your data mobility happens at a very good bandwidth all the applications are running uh properly so we normally used to have in the data center this kind of three tier approach previously but now we have moved on to the spine leaf architecture .

Why because for example now if i am uh sitting over here and i'm trying to put a request on a website so what happens is uh here i'm i'm sitting i put a request to the server okay now this server is basically going to be my web server let's say now what it is going to do it is being attached from the back end .

To the database okay so as i put a request it it fetches the data and then takes it and forwards it to the client now what happens is most of the time the traffic flow here what you see that the traffic flow from .

Was from south to north okay in order for this pc to communicate with this pc we had to go via such kind of approach or via this kind of approach any approach you take so it was moving from south to north approach in order to reach to this pc now in us data center where we see that .

Majority of our end devices are going to be servers okay as we see in our campus we have these computers or our host in a data center it will be majority majority of the times what you see what you will see that you will connect the servers or the databases or your van .

Network towards your like it will be servers are going to be your end devices i can simply say okay they are going to be your end devices so majorly the traffic flow is in the east west direction okay the traffic flow is not in the .

North south or south north direction i can simply say that the traffic flow okay if this server wants to communicate to this server so there is a east-west flow of traffic which was analyzed by the people that the traffic is mostly east-west direction so this architecture was developed with .

That in that regards that okay if any maximum if this server has to communicate to this server this is my application server this is my database server i can have any other server or dns server or hosting over here so if i have to speak i am majorly moving towards east-west .

Direction so maximum i and the architecture is so simple and this is known as your spine switches which is known as your backbone switches and these are known as your leaf switches so the connection is pretty simple that each leaf switches will have a dedicated .

Link towards the spine switches and you will wonder that okay you're also man you see some some sort of you know this orange link so again there will be a kind of bandwidth issue all your high bandwidth will not be available to you but we don't use the spanning tree .

Protocol the way i just mentioned over here that spanning tree protocol is used to make your uh entire network loop free because loops can make disaster in the network okay it can just make your it can just create disaster in your network so in order to avoid loops we had this spanning tree protocol now what are .

Loops i know some of you might be wondering ma'am what are what is she talking about loops so i'll discuss about that don't worry but just understand that when you are connecting the two devices and if there is a loop formed okay so there is going to be some issue and .

That has been avoided by the spanning tree protocol however in spine leaf architecture we are not dealing with the spanning tree protocol we deal with something else which is known as thrill okay now what is that again will be understood later as of now you just need to understand .

That in a spine leaf architecture we have most of the time the end devices are host uh sorry the end devices are servers and in order to connect to any other server i just need to do one hop i can just reach to reach this server i can just reach to this this .

This spine uh switch and then directly to the axis switch again from this i can so that is the latency involved would be very low and the scalability so see what are we interested in when we are designing a network what are the objectives which we have in .

Our mind the objective is that first and foremost we should have good bandwidth high bandwidth should be available second objective while designing the network is to have low latency okay the third objective which we generally have is the scalability that .

It should be easy to scale your network as per the requirements so if any server has to be added i just need to add one spine switch one access one lead switch and then directly make one connection from here to your and then it is ready to go okay it doesn't require a lot of .

Things to be done unlike a 380 architecture it is also good in terms of scalability however you can see the kind of connection mesh connections we have here so on on a counterpart the spine leaf is very much easier to scale okay so bandwidth .

Good bandwidth low latency scalability highly availability all these things you will see okay we need high available networks to be designed that the downtime is very low and apart from that we also need to ensure that we have maintaining good security we are .

Maintaining uh good quality of services all these things are one of the several five to six parameters which we see when we are designing a campus network or a data center network whatever we see okay so this architecture is being adopted now and it's being .

Followed because of the various advantages it has it is not using this fine tree uh stp protocol it uses another type of protocol okay so that all the links okay now you can see that there are four links which is three links which are going now from this switch seven the link here is .

Blocked here is blocked so now this was a case when we were designing your okay like a three tier architecture so only one link is active at this time though there was three links so it should have if each port is supporting or each link is giving me 100 mbps of speed and i could just see .

That two of them are blocked so i am instead of giving getting 300 i'm only getting 100 so that is known as to be very bad in terms of utilization of bandwidth okay the uplinks has to be very good in terms of support the uplink these are known as the uplink ports and their bandwidth is usually much higher than .

The normal switch ports okay the uplink ports are supposed to have a higher bandwidth than the other switch ports okay this is something again uh which i want to tell you okay so i hope you got the basic difference about the two tier three tier this is also known as a traditional .

Architecture okay the two tier is just collapsing the core layer and then this is the spine leaf architecture where you just you're in this architecture most of the time this is there in the the architecture is designed for a data center .

Where your end devices are generally the servers so since we have normally the traffic in the east-west direction so this architecture gives you good scalability it gives you good uh like the latency is very low and hence it is much preferred okay so i hope you people have got the understanding about .

The traditional architecture as and the sign leaf architecture all right so we have achieved that we have understood that in short now coming back to so we completed lan van man we completed on uh two tier three tier and spine leaf architecture now coming back to this on premises and .

Cloud topology because you're talking about all topologies on-premises and cloud yeah yeah uh i just wanted to know why east west the traffic flow why east-west the traffic flows okay so you can see from this uh diagram okay i'll show you here .

Okay you see that if i have to this is my application server let's see okay these are the three servers which are now there in my system let me clear the diagram okay so we see here that let's say this is my application server okay this one is my application server .

This is my database server so whenever a request comes from a user okay to access the application server and i have to fetch some data from the database initially people used to have different servers for different app different like for computation purpose .

For database for application we used to have different servers like so to fetch the data i have to move towards my towards this server am i right okay so your traffic is going from it it is move it has to it is it wants to move towards the west or towards the east .

Direction because your servers are being kept in that way they have been placed in that manner so your traffic is generally flowing it wants to access the servers which are placed around that so you will see that that that traffic mostly flows from in the east-west direction .

So that from your application server can access the database server some other server can actually access the some other server so in that manner we say it is east-west direction here in this architecture what you see that there was a latency involved because of these south to north .

Okay south to north way the traffic was flowing okay so based on that you see that south to north okay because of this approach the hierarchical threat air approach okay there was a high latency involved over there as comparison to spine leaf now since .

You see there is no interconnection between the spine switches there is no interconnection between the leaf switches so what they have to do is directly they go they can reach to one specific one of the spine switch and directly access to the other server which is sitting over .

There so there is not much of you know latency you would see here maximum one hope or to maximum one hop which we can see between the two access layer maximum one home however this was not the case with the three tier architecture okay so the eastwood direction of the .

Flow is because of the traffic flow because your servers are laid down in that specific manner or that arrangement i can say so that is why this architecture is much useful because even if this pc wants to speak to this pc you see that the direction is not .

In the horizontal or in a vertical manner okay these are helping us this devices networking devices are helping us reach our traffic but we are communicating let's say they are on the same layer we can say they are in the same layer layer wise if i say this is this is on .

The same layer they are on the same layer so accordingly the traffic if it is in the same layer we can say that east west is going to be much better approach than the south north or north south yes great thank you so yeah okay .

Any more question guys anyone has any questions to us here yeah can we use l2 as a distribution l2 at the distribution no that is not possible because l2 layer switches is only responsible for you switching the uh switching within .

The compute like within this network okay within this network how many ever host you are connecting this will help you out but if this okay somebody is sitting over here i cannot say they are in the same network right they now we are going to learn about vlan also so there they can be uh let's say .

These are all the members of your one of the department in your organization there are several departments in your organization so let's say this is my hr department they are all the hr people sitting here we have an accounts department who are sitting here .

Okay so if these two people one they want to communicate with each other okay definitely we can directly communicate if we if we have a very small network i can just connect the two switch switches and make use of vlan communication okay inter vlan communication that we are .

Going to learn okay they are in one vlan and this is in other vlan so in order to have an interview and communication we need to keep also have router in picture if you if you have any idea about the vlan i don't know if you have any idea about the vlan okay okay yeah .

We will be discussing about vlan so when you want to talk with different vlans you need you will require a router so as well as in in l3 okay when you take a l3 switches or multi-layer switches you can take features from both switching as well as routing we can connect it through the okay so we can connect it through like .

Interval and routing right right right right right so you need for an inter vlan communication you either require some router on stick method there are different methods for inter vlan communication or you just place a layer 3 switch and you know can achieve the job of doing interview and .

Communication spine lift architecture is it widely used right now in the industry based yes industrial wise site i guess right yeah wherever we find the servers okay in the server environment okay where your end devices are in that is why i mentioned data centers okay like in the data center environment we .

Generally use this this uh architecture however when you are developing a very small campus network then in that manner we either go for a two-tier architecture majority we go with the two-tier or three-tier hierarchical architecture okay okay thank you .

Okay so yeah so guys those are spine leaf which is east-west traffic topology which is also used in the aci data center model these days okay and main thing is a reduction in the hop counts here only there is one hop count between the servers and the code devices .

Whereas in the traditional there will be two or more hops so the packet reachability from end to end is very uh much less fine leaf okay yeah okay so i'll just take one last topic at present because uh .

I have got some work but one last topic i would cover here is the on-premises and cloud now these days we are hearing a lot about cloud and i'm sure you people must have almost some of you might be working in cloud technology or else or must have at least come .

Across this amazon web services aws cloud or microsoft anyone just let me know to chat how many of you have come across this cloud technology i guess most of you would have oh yeah right so what uh what is the difference like anybody who knows like .

Uh what what is the difference of having a on premise and a cloud technology let's say yeah azure so let's say for example you you you dream to start make your own startup okay and you want to develop some application you want to do something you want to .

Create your own business now when you want to create your own business you definitely need to have a good you know kind of let's say you want to develop a website for it or and many more thing you want to do it so initially what people used to do is if they want to create their own business or develop their own business .

They have to think from all dimensions right from developing their applications their softwares to the point of launching it and you know taking uh looking at it that how people could how my application is reachable to everyone so that way networking was one of the important component that you buy .

Your own servers you buy your own network routers your switches all cables then you buy the infrastructure to place all your devices and then you keep a team of people to maintain your devices configure do monitoring so lot of things was involved when you have .

To make your own business however with the cloud technology now they have taken all your burdening okay so they have just said that we have all the devices we have all the servers we have everything you just you can just use it on the rental basis okay you just use our whatever kind of .

Servers you require how many the way number of servers if you require one server to server or you require so much of memory for storage anything you just you know take it for rental purpose that is you just pay them for the time you utilize those .

Machines so you need not worry of going and purchasing it and keeping it for the sake of just if the things are not utilized well anyways you will get your business will collapse so just utilize it try on if you want if you're experimenting something don't worry about the infrastructure you .

Just concentrate on your application and we will look after whatever you want to wherever you whatever services like infrastructure you want we are going to give you on rental basis and you just focus on your applications so that way for many of the startups it has become very .

Much convenient that they don't have to bother about going and purchasing it and then you know developing everything from the scratch so that is the difference between on-premises and cloud though many of us uh in since all of the most of the companies have their on-premises .

Uh network which is already been established but still again they are trying to also move towards cloud okay they are making use of a hybrid architecture of making use of their environment on premises as well as developing some of their applications on cloud but most of the startups because they cannot take a headache of going and .

Purchasing so so expensive equipments and then keeping people for maintaining it and if it is not utilized properly then you will have to take extra attention so they just go on with the cloud services they just go and see their requirements and then deploy their applications so .

That is the difference between the on-premises and the cloud support okay so i end with this if anyone has any question i'll take up and tomorrow we'll discuss the other things which are there in line okay if anyone has any question okay good .

All good okay so just just summarize we just completed the smiley two-tier three-tier architecture we understood the difference between on-premises and cloud thing and the difference between lan man so tomorrow we'll start with the ipv4 addressing difference between .

Two tier and three tier in a two tier architecture you collapse your core okay your core and distribution they are known as a cd layer okay they act they are one and they are taking uh the features of both so core layer is not there only the access layer switches and then the cd switches we see the core distribution .

Layer switches that is a two tier and three tier where you have three tiers like access layer distribution layer and then the core layer so we saw in the left side the three tier architecture because we were having core switches if i remove the core switches .

I can simply say that it is a two tier architecture and that layer which is known as which was known before as distribution is now known as core distribution layer so for a small business environment or for a small campus network you go with the two tier and for .

The larger campus we go with the 3d okay guys so hope you enjoyed the lecture go and practice this try to make some scenarios like this and see how your connections happens definitely we are going to do most of the configurations for such kind of networks in our future classes okay okay all right good day good night .

Good night thank you bye good night good night


Most Popular