YT YouTube2Text

YouTube2Text → Transcripts → Network Layer

Network Layer — Transcript

by Internet and Web Technologies · 3,098 words · 607 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:00the network layer provides support for
0:03internetwork communication it
0:05facilitates packet forwarding and
0:07routing
0:07between devices on separate networks and
0:10enables end-to-end communication
0:13our discussion of the network layer will
0:15be split into two parts
0:16in this first lecture we will discuss
0:18the architecture of the internet
0:20how routers work the internet protocol
0:24and the structure of the ip packet
0:27in the following lecture we'll look into
0:29the various routing protocols which
0:31dictate the rules
0:32that routers and autonomous systems
0:34follow on how to forward packets such
0:36that they reach their final destination
0:38as efficiently as possible we will talk
0:41about
0:42shortest path algorithms like dijkstra
0:44and bellman ford
0:46and how they are used to route traffic
0:47on the internet
0:50the consequence of doing lectures in
0:51this order is that during this first
0:53lecture
0:54you have to assume that each router
0:56knows the best path to forward packets
0:58to
0:58and that this entire process is
1:00efficient
1:01more concisely if we are sending a
1:03message over the internet
1:04from computer a to b at each router we
1:08can use an oracle
1:09to determine which router will be the
1:11next hop
1:12the details about this oracle algorithm
1:14will be revealed in the next lecture
1:16as well as its runtime complexities
1:20a local area network or lan is a
1:23grouping of computers that are
1:24interconnected within a single physical
1:27network
1:28typically devices on a lan can easily
1:32communicate with each other
1:33but require special configuration in
1:35order to communicate with devices
1:37outside the local network networks
1:40are separated by routers a single
1:43network may have multiple routers
1:46each acting as a gateway between other
1:48networks
1:50if the topology of multiple connected
1:52networks is known beforehand
1:54and we have the mac addresses of all
1:56intermediate nodes
1:58this can be done solely through the
2:00datalink layer through the use of
2:01network bridges
2:03which are devices that connect two or
2:04more lands
2:07even with routers in play if we control
2:09the entire network
2:11we can communicate using just the data
2:13link layer
2:16when using the internet though we need
2:18to be able to communicate with networks
2:19that we don't own and are
2:22connected to routers that we have no
2:24control over
2:26the network layer facilitates delivery
2:28of these messages
2:30by getting us closer and closer to the
2:32final destination
2:33one hop at a time ipv4
2:37a network layer protocol facilitates
2:39communications between networks
2:41that are not known beforehand
2:45the most widely used protocol on the
2:47network layer is the internet protocol
2:50abbreviated ip it has two versions
2:54ipv4 and ipv6
2:58other protocols like icmp also exists
3:02but it's mainly used for error reporting
3:04purposes and its header is largely the
3:06same as the ip datagram
3:08with a few additional fields before your
3:11computer sends a message off to the
3:13internet
3:13it attaches an ip header to it this
3:16header for the most part stays the same
3:18at each hop along the path
3:20certain values like ttl are decreased
3:23but the address fields are maintained
3:25this is in contrast to data link layer
3:27protocols like ethernet
3:29and 802.11 where the frame headers are
3:32removed at each hop
3:34the ip header contains multiple fields
3:36the most important of which
3:38is the source and destination address
3:41addressing in ipv4 is done with ip
3:44addresses
3:45each device connected to the internet is
3:48identified by this
3:50these ip addresses are leased to
3:52individuals by internet service
3:54providers
3:55or isps your monthly internet bill
3:57includes payment for a temporary lease
4:00of an ip address from your isp's pool of
4:03available addresses
4:05you are not guaranteed the same ip
4:06address each time unless you pay for
4:08business class internet
4:10that includes a static ip
4:13ip addresses are a 32-bit number
4:15composed of four octets
4:17each octet is typically represented as a
4:20number from 0 to 255
4:22separated by dots for example
4:26172.6
4:31this format is called dotted decimal
4:34notation
4:36due to ip addresses being a 32-bit
4:38number
4:39there is an upper limit of 2 to the 32
4:42or roughly 4.2 billion addresses
4:45while enormous at the time of the
4:46internet's creation
4:48the wide adoption of it has caused these
4:51numbers to quickly deplete
4:53by january 2020 all ipv4 addresses
4:56have been allocated between various isps
4:59around the world
5:01this problem is called the ipv4 address
5:04exhaustion problem
5:07ip addresses are composed of two parts a
5:10network portion
5:11and a host portion the bits of an ip
5:14address
5:15designated for the network portion
5:18uniquely identifies the network
5:20while the host portion allows all
5:22devices on the network to share the same
5:24prefix
5:26by decreasing the size of the network
5:28portion you allocate more devices
5:30to a single host allowing them to have
5:33more
5:33connected devices
5:36by decreasing the number of bits in the
5:38networking portion
5:40you allocate more bits for the host
5:43components
5:44allowing a single network to have more
5:47devices
5:49the drawback to this is that you have
5:51fewer unique networks
5:53which exacerbates the ipv4 exhaustion
5:56problem
5:57especially if these hosts remain unused
6:00this next section is historical so we
6:02don't really use this nowadays
6:04but it helps us understand the ipe v4
6:07exhaustion problem better
6:09originally these network addresses were
6:11divided into network hosts
6:13at every eight bits or one octet
6:16creating three classes of addresses
6:19class a networks
6:20identified with a first octet in the
6:23range of 0 to 126
6:26had one octet for its network portion
6:29and 3 octets for the host portion
6:31this allowed for 2 to 24 minus 2
6:34or roughly 16 million devices per
6:37network
6:38class b addresses identified with the
6:41first octet
6:42in the range 128 to 191
6:48had two octets for the network portion
6:51and two octets for the hosts
6:54this allowed for 2 to the 16th minus 2
6:58or 65 534 devices per network
7:04finally we have class c addresses which
7:06are identified with
7:08the first octet in the range 192-223
7:14these had three octets for the network
7:16portion and one octet
7:18for the host this allowed for 2 to the 8
7:21minus 2 or 254 hosts
7:26the drawback to this classful system was
7:28because there were only three rigid
7:29sizes of ip address blocks
7:32small companies could outgrow their
7:33allocated class c address
7:35and their only option would be to either
7:37get another classy address
7:39and combine the two but this would
7:41create two separate ranges of ip
7:43addresses for a single organization
7:46their other option was to upgrade to a
7:48class b address
7:50giving them 256 times as much addressing
7:53space as they previously had
7:56this could also happen to organizations
7:58that outgrew their class b address
8:00but with the added complication that
8:03there could only ever exist
8:05126 class a addresses in modern
8:08applications
8:09we use classlist interdomain routing or
8:12cider
8:14insider instead of an entire octet of 8
8:17bits
8:17being changed at once a single bit is
8:20moved from the network to the hosts
8:23when using cidr ip addresses have two
8:26forms
8:27one to identify the network called the
8:29citer address
8:31this includes a network portion a host
8:33portion
8:34and a subnet identifier for example
8:3864.60
8:44the number after the slash represents
8:46the subnet identifier
8:48and specifies the number of bits that
8:51will be used for the
8:52network portion
8:56in this example 20 bits will be
8:58allocated for the network
9:01and the remaining 12 bits are used for
9:03hosts
9:05this network can have up to two to the
9:0712
9:08minus two hosts there are two reserved
9:11addresses for each network
9:14the host address of all zeros is
9:16reserved to identify the network itself
9:18the host address with all ones is
9:20reserved as a broadcast
9:22address there is also an ip address
9:25for individual hosts the key difference
9:29is the lack of a subnet identifier
9:32when sending messages to individual
9:35machines
9:36we do not care about how the ip
9:38addresses are allocated
9:40all we care about is that each address
9:43is unique
9:45so ip addresses are used to identify
9:48individual machines while cider
9:51addresses
9:52represent ranges of ip and are used to
9:55identify entire networks
9:57so ip addresses represent individual
10:00machines
10:02while cider addresses represent ranges
10:05of ip addresses and are used to identify
10:08entire networks slider fixes the
10:11allocation problem
10:12of classful domain routing networks
10:15looking to add more hosts
10:17could exchange their existing site
10:19address for an
10:20available address with one bit shifted
10:22over
10:24as an example a network could return
10:27its siter address 64.64.0.0.20.
10:32in exchange for the siter address
10:3680.80.0.0.19
10:38because the network portion has been
10:40reduced by a single bit
10:42this network can have twice as many
10:44hosts while ensuring all devices on the
10:46network share the same network prefix
10:49the protocol data unit of ipv4
10:52is the ipv4 packet also called the ipv4
10:56datagram
10:58this header wraps our payload and gets
11:01re-wrapped with a datalink layer header
11:03before being transmitted
11:06the first four bits represents the
11:09version
11:10and determines which version of ip this
11:12packet is using
11:14the field after that is the internet
11:16header length
11:18the size of the ip header is variable
11:22because of the options fields
11:26there is a minimum header size of 20
11:27bytes which is also the most common
11:29header size
11:31we can derive that value by noticing
11:34that
11:34in this diagram each row is 32 bits
11:38or 4 bytes in length and that there are
11:42five rows not including the options
11:44fields
11:46the next field type of service specifies
11:48data about
11:49quality of service technologies the
11:52field after that
11:52total length specifies the length of the
11:55entire packets
11:56the maximum size of a ip packet is 65
12:00535 bytes the next row
12:04deals with the fragmentation of packets
12:06should a particular packet
12:08exceed the maximum transmission units or
12:10mtu of a link
12:12that packet will need to be fragmented
12:14before it is sent off
12:16packets are reassembled at their final
12:18destination
12:19the identification field which is a 16
12:22bit fields
12:23is used to collect and group together
12:27fragmented messages when ipmessages are
12:30fragmented
12:31this field is used to identify which
12:34fragments
12:35belong to the same original message
12:38there are three flags that follow the
12:41first bit
12:42is a reserved flag and is always zero
12:45the second bit is the don't fragment bit
12:48and indicates if a datagram is allowed
12:50to be fragmented
12:52the last bit is the more fragments
12:56flag and indicates if there are more
12:58fragments
13:00the next field the fragment offsets is a
13:0313-bit field
13:05used to help sequence packets that are
13:07fragmented
13:09and it does so by indicating to the
13:11recipient's device
13:13where in the overall message each
13:15particular fragment should be placed
13:18moving on to the next row the first
13:21field that we will talk about
13:22is the time to live or ttl
13:25this is a 8-bit field and it represents
13:28how many router hops
13:30a datagram can traverse before it's
13:32discarded
13:34we discard long-lived messages on the
13:36network
13:37to prevent loops when there is a
13:40misconfigured
13:41or malicious router without the ttl
13:44it is possible for a misconfigured
13:46router to cause a packet to stay on the
13:48wire forever
13:49forwarding messages such that it
13:50eventually comes back to itself
13:52the ttl decrements after each hop and
13:55when they hit 0
13:57the packet is dropped the default ttl
14:00is 64 hops next the protocol
14:04is a 8-bit field used to determine the
14:07transport layer protocol
14:08use one layer above the next field is
14:11the header checksum
14:13which is a lesser use 16-bit field
14:16used for data integrity the checksum is
14:20not really used
14:21because both major datalink layer
14:23protocols
14:24ethernet and 802.11 provide
14:28data integrity using the frame check
14:31sequence
14:32if we look at the ipv6 header field in
14:35fact
14:35we will see no checksum fields and in
14:38fact there is no
14:39data integrity field instead that task
14:42is offloaded to the data link layer
14:45when a router receives an ipa packet
14:47first we'll see if the incoming package
14:49destination address
14:50is for a host on its own network and if
14:52it is the router makes the final
14:54delivery
14:55otherwise if the destination address
14:58matches a network cider range in the
15:00routing table
15:01it will forward the message to the next
15:03router out on a specified interface
15:06we will go into the details of how the
15:08routing table is populated with entries
15:10in our next lecture all routers are
15:13configured with a special catch-all
15:15entry called the default gateway
15:17typically your isp which is an ip
15:20address to forward messages to
15:21in the event that no other forwarding
15:23rules apply
15:25so let's look at a sample exchange
15:26between two hosts
15:28host one and host two when arriving on
15:31the router
15:32the packet will have its ttl field
15:34decreased by one
15:36and the entire packet will be dropped if
15:38the ttl field hits zero
15:41then the router looks up the next hop in
15:44its own routing table
15:46and performs fragmentation if required
15:49before sending the data
15:51to the next hop out on a specified
15:54interface these individual hops
15:57are chained together until we reach the
16:00final destination
16:02routers are considered network layer
16:04hardware
16:05because they inspect the destination ip
16:08address in each packet
16:09before making a decision on what
16:12interface
16:13to forward the packet to there are
16:15several ip ranges
16:17that are not to be allocated to any
16:18entity
16:20these addresses are meant only to be
16:23used locally
16:25traffic using these reserved ip ranges
16:27never leaves the network
16:30there are many reserved ip ranges but
16:32for this course
16:33we only care about three 127.0.0.08
16:38block
16:41this is used for the loopback address
16:44which is when a host wants to send a
16:47message
16:48to itself it is used quite often
16:51in net code development as the operating
16:54system
16:54will simulate the requests without
16:57needing any actual hardware
16:59the siter address
17:02slash 192.0.0.0.24 is used for
17:06a local area network it can support up
17:08to 256 devices
17:11if this is not enough you can also use
17:13the 10.0.0.0.8 range but we won't be
17:18discussing that
17:20finally the range 224.0.0.0.4
17:27is used for ip multicast
17:30ip multicasting uses a one-to-many
17:33association where datagrams are routed
17:35simultaneously
17:37in a single transmission to multiple
17:38recipients
17:40a server only needs to send the message
17:42once
17:43and machines will listen in on a
17:45specified ip
17:47ip multicasting uses a one to many
17:49association
17:50where datagrams are routed
17:51simultaneously in a single transmission
17:54to multiple recipients a server only
17:57needs to send the message
17:58once and all configured machines
18:01will listen in on the specified ip
18:04video surveillance systems are probably
18:06the most common usage
18:08with many closed circuit tvs
18:10broadcasting on specific ip addresses
18:13that you can navigate your browser to
18:15and watch
18:17because these are local ip addresses
18:19multicast traffic
18:20will never leave your network other
18:23usages
18:23include pc hosting and software updates
18:26voiceover ip video monitoring and
18:29conference calls
18:31your home network can have up to 256
18:34devices
18:35in the address block 192.0.0.0.8.
18:40when a device connects to your home
18:41network it has temporarily
18:43allocated one of these addresses through
18:45a process called dhcp
18:48the dynamic host configuration protocol
18:51is used to
18:52temporarily assign a local ip address
18:55to a newly connected device
18:59the dhcp protocol has four stages
19:02dhcp discover where a newly connected
19:05device
19:06sends a discovery request that is
19:08broadcasted on the entire network
19:10and includes the device's mac address
19:13dhcp
19:14offer all dhp servers who receive a
19:17request
19:18and have an available ip address offered
19:20to lease the ip address to the newly
19:22connected device
19:24dhcp requests the newly connected device
19:28sends a request to the server of the
19:29first response they receive
19:32when receiving multiple offers it
19:34assumes that the first offer
19:36has the least latency dhcp
19:39acknowledgement
19:40upon receiving dhcp requests the server
19:44creates a record of it in the dhcp table
19:48to keep track of which local ip
19:50addresses are currently in use
19:53then it sends an acknowledgement which
19:55includes additional information
19:57like expiration ipaddresses assigned
20:01with the dhcp
20:02protocol are designated to be dynamic
20:04ips
20:06dynamic ips revert back to the dhcp
20:08server
20:09once they are expired
20:13they are useful for devices that don't
20:15need to communicate with other local
20:17devices
20:18like desktops and laptops
20:21we can also go to the router settings to
20:23configure a static ip address
20:25unlike dhcp static ip addressing does
20:28not expire unless revoked
20:30they are useful for devices that others
20:33may want to connect to
20:34like a network printer this is because
20:36dhcp will automatically assign any newly
20:39connected device
20:40an ip address from its available pool if
20:43a lease expires
20:44or we lose power there is no guarantee
20:47that the devices when they reconnect
20:49will get the same ip address
20:52when using a network printer we are
20:54sending a print job to an ip address
20:57should the ip address of that printer
20:59change because dhcp gave it a new one we
21:02may wind up sending our print jobs to
21:03the wrong location
21:05ip addresses belong to the network not
21:08the devices attached to those networks
21:11these devices lease an ip address using
21:13the dhcp protocol this is in contrast to
21:17the mac address
21:18which stays the same regardless of where
21:21you are connected to
21:24each computer connected to a network is
21:26least a local ip address
21:28this address is unique to the network
21:31but
21:31can be reused outside the network as it
21:34is a reserved
21:35address which will never leave the local
21:37network
21:38finally we'll discuss the address
21:40resolution protocol
21:41or arp arp
21:44is a data link layer protocol used to
21:47map a known ipv4 address
21:50to an unknown mac address it cannot be
21:53used
21:53outside of local networks it is used
21:56predominantly to
21:58find a mac address given a ip address
22:03all network devices will keep an arp
22:06table
22:06a list of ip addresses and known mac
22:08addresses associated with each device
22:11arp table entries expire after a short
22:14period of time
22:15to ensure network changes are accounted
22:16for
22:18in a windows machine you can look at
22:20your arp table
22:21by typing in to a command prompt arp
22:25a if a router receives a
22:28unknown ip address it will broadcast an
22:32arp
22:32request out on the entire network
22:36the device that owns that ip address
22:39will respond back with its mac address
22:42after which the router will update its
22:46arp table before forwarding the
22:49request in this lecture there was a key
22:52assumption that we made
22:54that our routers already knew the best
22:56path
22:57to get our message delivered in our next
22:59lecture
23:00we'll discuss routing protocols which
23:02populate our router's memory
23:04called a routing table with cider
23:06addresses
23:07and how it keeps those records up to
23:09date with a constantly changing internet

About this transcript

This page contains the full transcript of Network Layer by Internet and Web Technologies, generated from the public captions YouTube serves with the video. The transcript has 3,098 words across 607 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 choose a caption track available on the video.

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, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.