Link State Routing (OSPF)

Link-state routing is the second major class of intradomain routing protocol. The basic idea behind link-state protocols is very simple: Every node knows how to reach its directly connected neighbors, and if we make sure that the totality of this knowledge is disseminated to every node, then every node will have enough knowledge of the network to build a complete map of the network. This is clearly a sufficient condition (although not a necessary one) for finding the shortest path to any point in the network.

link-state routing protocols rely on two mechanisms: reliable dissemination of link-state information, and the calculation of routes from the sum of all the accumulated link-state knowledge.

Reliable Flooding

Reliable flooding is the process of making sure that all the nodes participating in the routing protocol get a copy of the link-state information from all the other nodes. As the term “flooding” suggests, the basic idea is for a node to send its link-state information out on all of its directly connected links, with each node that receives this information forwarding it out on all of its links. This process continues until the information has reached all the nodes in the network.

each node creates an update packet, also called a link-state packet (LSP), that contains the following information:

• the ID of the node that created the LSP
• a list of directly connected neighbors of that node, with the cost of the link to each one
• a sequence number
• a time to live for this packet

Consider a node X that receives a copy of an LSP that originated at some other node Y. Note that Y may be any other router in the same routing domain as X. X checks to see if it has already stored a copy of an LSP from Y. If not, it stores the LSP. If it already has a copy, it compares the sequence numbers; if the new LSP has a larger sequence number, it is assumed to be the more recent, and that LSP is stored, replacing the old one. A smaller (or equal) sequence number would imply an LSP older (or not newer) than the one stored, so it would be discarded and no further action would be needed. If the received LSP was the newer one, X then sends a copy of that LSP to all of its neighbors except the neighbor from which the LSP was just received. The fact that the LSP is not sent back to the node from which it was received helps to bring an end to the flooding of an LSP. Since X passes the LSP on to all its neighbors, who then turn around and do the same thing, the most recent copy of the LSP eventually reaches all nodes.

Each node generates LSPs under two circumstances. Either the expiry of a periodic timer or a change in topology can cause a node to generate a new LSP. However, the only topology-based reason for a node to generate an LSP is if one of its directly connected links or immediate neighbors has gone down. The failure of a link can be detected in some cases by the link-layer protocol. The demise of a neighbor or loss of connectivity to that neighbor can be detected using periodic “hello” packets. Each node sends these to its immediate neighbors at defined intervals. If a sufficiently long time passes without receipt of a “hello” from a neighbor, the link to that neighbor will be declared down, and a new LSP will be generated to reflect this fact. One of the important design goals of a link-state protocol’s flooding mechanism is that the newest information must be flooded to all nodes as quickly as possible, while old information must be removed from the network and not allowed to circulate.

Route Calculation

Once a given node has a copy of the LSP from every other node, it is able to compute a complete map for the topology of the network, and from this map it is able to decide the best route to each destination. The question, then, is exactly how it calculates routes from this information. The solution is based on a well-known algorithm from graph theory—Dijkstra’s shortest-path algorithm.

The algorithm is defined as follows:

M = {s}
for each n in N− {s}
C(n) = l(s, n)
while (N = M)
M = M ∪ {w} such that C(w) is the minimum for all w in (N− M)
for each n in (N− M)
C(n) = MIN(C(n), C(w) + l(w, n))

Each switch computes its routing table directly from the LSPs it has collected using a realization of Dijkstra’s algorithm called the forward search algorithm. Specifically, each switch maintains two lists, known as Tentative and Confirmed. Each of these lists contains a set of entries of the form (Destination, Cost, NextHop).

The algorithm works as follows:

1. Initialize the Confirmed list with an entry for myself; this entry has a cost of 0.
2. For the node just added to the Confirmed list in the previous step, call it node Next, select its LSP.
3. For each neighbor (Neighbor) of Next, calculate the cost (Cost) to reach this Neighbor as the sum of the cost from myself to Next and from Next to Neighbor. 
• If Neighbor is currently on neither the Confirmed nor the Tentative list, then add (Neighbor, Cost, NextHop) to the Tentative list, where NextHop is the direction I go to reach Next. 
• If Neighbor is currently on the Tentative list, and the Cost is less than the currently listed cost for Neighbor, then replace the current entry with (Neighbor, Cost, NextHop), where NextHop is the direction I go to reach Next.
4. If the Tentative list is empty, stop. Otherwise, pick the entry from the Tentative list with the lowest cost, move it to the Confirmed list, and return to step 2.

The link-state routing algorithm has many nice properties: It has been proven to stabilize quickly, it does not generate much traffic, and it responds rapidly to topology changes or node failures. On the downside, the amount of information stored at each node (one LSP for every other node in the network) can be quite large.

The Open Shortest Path First Protocol (OSPF)

One of the most widely used link-state routing protocols is OSPF. The first word, “Open,” refers to the fact that it is an open, nonproprietary standard, created under the auspices of the IETF. The “SPF” part comes from an alternative name for linkstate routing.

 

LS age - The time, in seconds, since the LSA was generated. 
LSID (Link State ID) - The ID of the router that generated the LSA. 
Advertising Router - ID of the router that originated the LSA.
LS Seq (Link State Sequence) - The sequence number of the advertisement. Used to detect old or duplicate link state advertisements.  
Flags - Possible values:

• V - Router is the endpoint of an active virtual link that is using the area as a transit area. 
• ASBR - Router is an autonomous system boundary router (ASBR). 
• ABR - Router is an area border router (ABR).

 Link ID - Identifies the object to which this router link connects for each Link Type. Possible values:

•     If Link Type is PTP, then this is the neighboring router's router ID.
•     If Link Type is Transit, then this is the address of the designated router.
•     If Link Type is Stub, then this is the IP network or subnetwork number.
•     If Link Type is Virtual Link, then this is the neighboring router's router ID.

Link Data - Provides additional link information. Possible values:

•     If Link Type is PTP, then this is the MIB II index value for an unnumbered point-to-point interface.
•     If Link Type is Transit, then this is the IP address of the advertising router's interface.
•     If Link Type is Stub, then this is the network's IP address mask.
•     If Link Type is Virtual Link, then this is the IP address mask of the neighboring router.

Link Type - A description of the router link. Possible values:

•    PTP - Connection is point-to-point to another router.
•    Transit - Connection is to a transit network.
•    Stub - Connection to a stub network.
•    Virtual link - Connection is to a far-end router that is the endpoint of a virtual link.

Metric - Cost of using this outbound router link. With the exception of stub networks, this value must be other than 0. 

 

NUM_TOS :TOS information is present to allow OSPF to choose different routes for IP packets based on the value in their TOS field.

Distance-Vector Routing

Each node constructs a one-dimensional array containing the "distances"(costs) to all other nodes and distributes that vector to its immediate neighbors.

1. The starting assumption for distance-vector routing is that each node knows the cost of the link to each of its directly connected neighbors.
2. A link that is down is assigned an infinite cost.

To see how a distance-vector routing algorithm works, it is easiest to consider an example 

 

Distance-vector routing: an example network.

In this example, the cost of each link is set to 1, so that a least-cost path is simply the one with the fewest hops. We can represent each node’s knowledge about the distances to all other nodes as a table like the one given in Table. Note that each node only knows the information in one row of the table. The global view that is presented here is not available at any single point in the network.

Table 1: Initial distances stored at each node (global view).

We may consider each row in Table 1 as a list of distances from one node to all other nodes, representing the current beliefs of that node. Initially, each node sets a cost of 1 to its directly connected neighbors and ∞ to all other nodes. Thus, A initially believes that it can reach B in one hop and that D is unreachable. The routing table stored at A reflects this set of beliefs and includes the name of the next hop that A would use to reach any reachable node. 

Table2: Initial routing table at node A.

Initially, then, A’s routing table would look like Table 2.The next step in distance-vector routing is that every node sends a message to its directly connected neighbors containing its personal list of distances.

Table 3: Final routing table at node A.

For example, node F tells node A that it can reach node G at a cost of 1; A also knows it can reach F at a cost of 1, so it adds these costs to get the cost of reaching G by means of F. This total cost of 2 is less than the current cost of infinity, so A records that it can reach G at a cost of 2 by going through F. Similarly, A learns from C that D can be reached from C at a cost of 1; it adds this to the cost of reaching C (1) and decides that D can be reached via C at a cost of 2, which is better than the old cost of infinity. At the same time, A learns from C that B can be reached from C at a cost of 1, so it concludes that the cost of reaching B via C is 2. Since this is worse than the current cost of reaching B (1), this new information is ignored.

Node A can update its routing table with costs and next hops for all nodes in the network and  The result is shown in Table 3

If there is any topology changes, it only takes a few exchanges of information between neighbors before each node has a complete routing table. The process of getting consistent routing information to all the nodes is called convergence.

There are two different circumstances under which a given node decides to send a routing update to its neighbors.

One of these circumstances is the periodic update. each node automatically sends an update message every time interval, even nothing has changed.The frequency of these periodic updates varies from protocol to protocol.

The second mechanism, sometimes called a triggered update, happens whenever a node receives an update from one of its neighbors that causes it to change one of the routes in its routing table. That is, whenever a node’s routing table changes, it sends an update to its neighbors, which may lead to a change in their tables, causing them to send an update to their neighbors.

Final distances stored at each node (global view).

When a node detects a link failure

• F detects that link to G has failed
• F sets distance to G to infinity and sends update to A
• A sets distance to G to infinity since it uses F to reach G
• A receives periodic update from C with 2-hop path to G
• A sets distance to G to 3 and sends update to F
• F decides it can reach G in 4 hops via A

Slightly different circumstances can prevent the network from stabilizing.Suppose the link from A to E goes down. In the next round of updates, A advertises a distance of infinity to E, but B and C advertise a distance of 2 to E. Depending on the exact timing of events, the following might happen.

• Node B, upon hearing that E can be reached in 2 hops from C, concludes that it can reach E in 3 hops and advertises this to A
• Node A concludes that it can reach E in 4 hops and advertises this to C
• Node C concludes that it can reach E in 5 hops; and so on.
• This cycle stops only when the distances reach some number that is large enough to be considered infinite
• This problem is called as Count-to-infinity problem

 There are several partial solutions to this problem. The first one is to use some relatively small number as an approximation of infinity. For example, we might decide that the maximum number of hops to get across a certain network is never going to be more than 16, and so we could pick 16 as the value that represents infinity. This at least bounds the amount of time that it takes to count to infinity.

One technique to improve the time to stabilize routing is called split horizon. The idea is that when a node sends a routing update to its neighbors, it does not send those routes it learned from each neighbor back to that neighbor.For example, if B has the route (E, 2, A) in its table, then it knows it must have learned this route from A, and so whenever B sends a routing update to A, it does not include the route (E, 2) in that update.In a stronger variation of split horizon, called split horizon with poison reverse, B actually sends that route back to A, but it puts negative information in the route to ensure that A will not eventually use B to get to E.

Routing Information Protocol (RIP)

One of the most widely used routing protocols in IP networks is the Routing Information Protocol (RIP). RIP is the canonical example of a routing protocol built on the distance-vector algorithm.

RIP Packet Format

RIP is in fact a fairly straightforward implementation of distance-vector routing. Routers running RIP send their advertisements every 30 seconds; a router also sends an update message whenever an update from another router causes it to change its routing table. One point of interest is that it supports multiple address families, not just IP. The network-address part of the advertisements is actually represented as a family, address pair. RIP version 2 (RIPv2) also has some features related to scalability

Dynamic Host Configuration Protocol (DHCP)

We observed that Ethernet addresses are configured into the network adaptor by the manufacturer, and this process is managed in such a way to ensure that these addresses are globally unique. This is clearly a sufficient condition to ensure that any collection of hosts connected to a single Ethernet (including an extended LAN) will have unique addresses. Furthermore, uniqueness is all we ask of Ethernet addresses.

IP addresses, by contrast, not only must be unique on a given internetwork, but also must reflect the structure of the internetwork. They contain a network part and a host part, and the network part must be the same for all hosts on the same network. It is not possible for the IP address to be configured once into a host when it is manufactured, since that would imply that the manufacturer knew which hosts were going to end up on which networks, and it would mean that a host, once connected to one network, could never move to another. For this reason, IP addresses need to be reconfigurable.

Most host operating systems provide a way for a system administrator, or even a user, to manually configure the IP information needed by a host. However, there are some obvious drawbacks to such manual configuration.

It is simply a lot of work to configure all the hosts in a large network directly, especially when you consider that such hosts are not reachable over a network until they are configured. Even more importantly, the configuration process is very error-prone, since it is necessary to ensure that every host gets the correct network number and that no two hosts receive the same IP address. For these reasons, automated configuration methods are required. The primary method uses a protocol known as the Dynamic Host Configuration Protocol (DHCP).

DHCP relies on the existence of a DHCP server that is responsible for providing configuration information to hosts. There is at least one DHCP server for an administrative domain. At the simplest level, the DHCP server can function just as a centralized repository for host configuration information.

A more sophisticated use of DHCP saves the network admininstrator from even having to assign addresses to individual hosts. In this model, the DHCP server maintains a pool of available addresses that it hands out to hosts on demand. This considerably reduces the amount of configuration an administrator must do, since now it is only necessary to allocate a range of IP addresses (all with the same network number) to each network.

Since the goal of DHCP is to minimize the amount of manual configuration required for a host to function, it would rather defeat the purpose if each host had to be configured with the address of a DHCP server. Thus, the first problem faced by DHCP is that of server discovery.

To contact a DHCP server, a newly booted or attached host sends a DHCPDISCOVER message to a special IP address (255.255.255.255) that is an IP broadcast address. This means it will be received by all hosts and routers on that network. (Routers do not forward such packets onto other networks, preventing broadcast to the entire Internet.) In the simplest case, one of these nodes is the DHCP server for the network. The server would then reply to the host that generated the discovery message (all the other nodes would ignore it). However, it is not really desirable to require one DHCP server on every network because this still creates a potentially large number of servers that need to be correctly and consistently configured. Thus, DHCP uses theconcept of a relay agent. There is at least one relay agent on each network, and it is configured with just one piece of information: the IP address of the DHCP server. When a relay agent receives a DHCPDISCOVER message, it unicasts it to the DHCP server and awaits the response, which it will then send back to the requesting client. The process of relaying a message from a host to a remote DHCP server is shown

A DHCP relay agent receives a broadcast DHCPDISCOVER message from a host and sends a unicast DHCPDISCOVER message to the DHCP server.

Figure shows the format of a DHCP message. The message is actually sent using a protocol called UDP (the User Datagram Protocol) that runs over IP. 

DHCP PACKET FORMAT

DHCP is derived from an earlier protocol called BOOTP, and some of the packet fields are thus not strictly relevant to host configuration. When trying to obtain configuration information, the client puts its hardware address (e.g., its Ethernet address) in the chaddr field. The DHCP server replies by filling in the yiaddr (“your” IP address) field and sending it to the client. Other information such as the default router to be used by this client can be included in the options field.

Internet Protocol (IP)

Simple Internetworking

Internetwork refer to an arbitrary collection of networks interconnected to provide some sort of host to host packet delivery service.

An internetwork is often referred to as a “network of networks” because it is made up of lots of smaller networks. we see Ethernets, an FDDI ring, and a point-to-point link. Each of these is a single-technology network. The nodes that interconnect the networks are called routers. They are also sometimes called gateways, but since this term has several other connotations, we restrict our usage to router.

Internet Protocol

The Internet Protocol is the key tool used today to build scalable, heterogeneous internetworks. It was originally known as the Kahn-Cerf protocol after its inventors. One way to think of IP is that it runs on all the nodes (both hosts and routers) in a collection of networks and defines the infrastructure that allows these nodes and networks to function as a single logical internetwork.

Service Model

The IP service model can be thought of as having two parts: an addressing scheme, which provides a way to identify all hosts in the internetwork, and a datagram (connectionless) model of data delivery. This service model is sometimes called best effort because, although IP makes every effort to deliver datagrams, it makes no guarantees.

Datagram Delivery

The IP datagram is fundamental to the Internet Protocol. A datagram is a type of packet that happens to be sent in a connectionless manner over a network. Every datagram carries enough information to let the network forward the packet to its correct destination;The “best-effort” part means that if something goes wrong and the packet gets lost, corrupted, misdelivered, or in any way fails to reach its intended destination, the network does nothing—it made its best effort, and that is all it has to do. It does not make any attempt to recover from the failure. This is sometimes called an unreliable service.

Best-effort, connectionless service is about the simplest service you could ask for from an internetwork, and this is a great strength. For example, if you provide best effort service over a network that provides a reliable service, then that’s fine—you end up with a best-effort service that just happens to always deliver the packets. If, on the other hand, you had a reliable service model over an unreliable network

Packet Format

Version: Version no. of Internet Protocol used (e.g. IPv4)

Hlen: specifies the length of the header in 32-bit words. When there are no options, which is most of the time, the header is 5 words (20 bytes) long.

TOS: The basic function of TOS is to allow packets to be treated differently based on application needs.

Length: The length includes everything in the datagram – both header and the data, the maximum length is 65,535 bytes.

Identification : The identification field is needed to allow the destination host to determine which datagram a newly arrived fragment belongs to. All the fragments of a datagram contain the same identification value.

Flags: The flags field is of 3-bits, one is unused and two 1-bit fields are DF, MF.

DF means Don’t Fragment , its an order to the router not to fragment the datagram because the receiver is incapable of putting the pieces back to gether again.

MF stands for More Fragments. All fragment except the last one have this bit set. It is needed to know when all fragments of a datagram have arrived.

Offset:The Fragment offset tells where in the current datagram this fragment belongs. All fragments except the last one in a datagram must be a multiple of 8 bytes, the elementary fragment unit. Since 13 bits are provided, there is a maximum of 8192 fragments per datagram, giving a maximum datagram length of 65,536 bytes, one more than the Total length field.

Time to live: The Time to live field is a counter used to limit packet lifetimes. It is supposed to count time in seconds, allowing a maximum lifetime of 255 sec. It must be decremented on each hop and is supposed to be decremented multiple times when queued for a long time in a router. In
practice, it just counts hops. When it hits zero, the packet is discarded and a warning packet is sent back to the source host.

Protocol: When the network layer has assembled a complete datagram, it needs to know what to do with it. The Protocol field tells it which transport process to give it to. TCP is one possibility, but so are UDP and some others.

Checksum: The Header checksum verifies the header only. Such a checksum is useful for detecting errors generated by bad memory words inside a router.

Source address and Destination address: The Source address and Destination address indicate the network number and host number.

Option	Description
Security	Specifies how secret the datagram is
Strict source routing	Gives the complete path to be followed
Loose source routing	Gives a list of router not to be missed
Record route	Makes each router append its IP address
Timestamp	Makes each router append its address and timestamp

IP Addresses

Every host and router on the Internet has an IP address, which encodes its network number and host number. The combination is unique: in principle, no two machines on the Internet have the same IP address. All IP addresses are 32 bits long and are used in the Source address and Destination address fields of IP packets. It is important to note that an IP address does not actually refer to a host. It really refers to a network interface, so if a host is on two networks, it must have two IP addresses. However, in practice, most hosts are on one network and thus have one IP address.

IP addresses were divided into the five categories.This allocation has come to be called classful addressing.

The class A, B, C, and D formats allow for up to 128 networks with 16 million hosts each, 16,384 networks with up to 64K hosts, and 2 million networks (e.g., LANs) with up to 256 hosts each (although a few of these are special). Also supported is multicast, in which a datagram is directed to multiple hosts. Addresses beginning with 1111 are reserved for future use. Over 500,000 networks are now connected to the Internet, and the number grows every year. Network numbers are managed by a nonprofit corporation called ICANN (Internet Corporation for Assigned Names and Numbers) to avoid conflicts.

Datagram Forwarding in IP

IP routers forward datagrams in an internetwork. That forwarding is the process of taking a packet from an input and sending it out on the appropriate output, while routing is the process of building up the tables that allow the correct output for a packet to be determined.

The main points to bear in mind as we discuss the forwarding of IP datagrams are the following:

• Every IP datagram contains the IP address of the destination host. 
• The “network part” of an IP address uniquely identifies a single physical network that is part of the larger Internet.
• All hosts and routers that share the same network part of their address are connected to the same physical network and can thus communicate with each other by sending frames over that network.
• Every physical network that is part of the Internet has at least one router that,by definition, is also connected to at least one other physical network; this router can exchange packets with hosts or routers on either network.

Forwarding Strategy: 

• every datagram contains destination's address
• if directly connected to destination network, then forward to host
• if not directly connected to destination network, then forward to some router
• forwarding table maps network number into next hop
• each host has a default router
• each router maintains a forwarding table

Datagram forwarding algorithm

if (NetworkNum of destination = NetworkNum of one of my interfaces) then
deliver packet to destination over that interface
else
if (NetworkNum of destination is in my forwarding table) then
deliver packet to NextHop router
else
deliver packet to default router

For a host with only one interface and only a default router in its forwarding table, this simplifies to

if (NetworkNum of destination = my NetworkNum) then
deliver packet to destination directly
else
deliver packet to default router

Switching and Forwarding

nA network switch (sometimes known as a switching hub) is a computer networking device that is used to logically connect devices together on a computer network, by using a form of packet switching to forward data to the destination device.

A network switch is a multi-port network bridge that processes and forwards frames at the data link layer (layer 2) of the OSI model. Switches can be designed to route packets in addition to performing packet switching; these switches are commonly known as layer-3 or multilayer switches.

A switch adds the star topology to the point-to-point link,bus (Ethernet), and ring (802.5 and FDDI) topologies.

• Even though a switch has a fixed number of inputs and outputs, which limits the number of hosts that can be connected to a single switch, large networks can be built by interconnecting a number of switches.
• We can connect switches to each other and to hosts using point-to-point links, which typically means that we can build networks of large geographic scope.
• Adding a new host to the network by connecting it to a switch does not necessarily mean that the hosts already connected will get worse performance from the network.

A switch is connected to a set of links and, for each of these links, runs the appropriate data link protocol to communicate with the node at the other end of the link. A switch’s primary job is to receive incoming packets on one of its links and to transmit them on some other link. This function is sometimes referred to as either switching or forwarding.

How does the switch decide which output port to place each packet on? The general answer is that it looks at the header of the packet for an identifier that it uses to make the decision.

There are two common approaches

• datagram or connectionless approach
• virtual circuit or connection-oriented approach

A third approach, source routing, is less common than these other two, but it is simple to explain and does have some useful applications. 

Datagrams

The idea behind datagrams is incredibly simple: every packet contains enough information to enable any switch to decide how to get it to its destination. To decide how to forward a packet, a switch consults a forwarding table (sometimes called a routing table)

                                        Datagram forwarding: an example network

Connectionless (datagram) networks have the following characteristics:

• A host can send a packet anywhere at any time, since any packet that turns up at a switch can be immediately forwarded.This contrasts with most connection-oriented networks, in which some “connection state” needs to be established before the first data packet is sent.

• When a host sends a packet, it has no way of knowing if the network is capable of delivering it or if the destination host is even up and running.

• Each packet is forwarded independently of previous packets that might have been sent to the same destination. Thus, two successive packets from host A to host B may follow completely different paths (perhaps because of a change in the forwarding table at some switch in the network).

• A switch or link failure might not have any serious effect on communication if it is possible to find an alternate route around the failure and to update the forwarding table accordingly.

Virtual Circuit Switching

• nWidely used technique for packet switching
• nUses the concept of virtual circuit (VC)
• nAlso called a connection-oriented model
• nFirst set up a virtual connection from the source host to the destination host and then send the data

Where host A again wants to send packets to host B.We can think of this as a two-stage process. The first stage is “connection setup.” The second is data transfer.

In the connection setup phase, it is necessary to establish “connection state” in each of the switches between the source and destination hosts. The connection state for a single connection consists of an entry in a “VC table” in each switch through which the connection passes. One entry in the VC table on a single switch contains

• a virtual circuit identifier (VCI) that uniquely identifies the connection at this switch and that will be carried inside the header of the packets that belong to this connection.
• an incoming interface on which packets for this VC arrive at the switch.
• an outgoing interface in which packets for this VC leave the switch.
• a potentially different VCI that will be used for outgoing packets.

The semantics of one such entry is as follows: If a packet arrives on the designated incoming interface and that packet contains the designated VCI value in its header, then that packet should be sent out the specified outgoing interface with the specified outgoing VCI value first having been placed in its header.

The combination of the VCI of packets as they are received at the switch and the interface on which they are received uniquely identifies the virtual connection. There may of be many virtual connections established in the switch at one time. 

The incoming and outgoing VCI values are generally not the same. The VCI is not a globally significant identifier for the connection; rather, it has significance only on a given link—that is, it has link local scope.
 

Whenever a new connection is created, we need to assign a new VCI for that connection on each link that the connection will traverse. We also need to ensure that the chosen VCI on a given link is not currently in use on that link by some existing connection.

There are two broad classes of approach to establishing connection state. One is to have a network administrator configure the state, in which case the virtual circuit is “permanent.” Of course, it can also be deleted by the administrator, so a permanent virtual circuit (PVC) might best be thought of as a long-lived or administratively configured VC. The next a host can send messages into the network to cause the state to be established. This is called as signalling, and the resulting virtual circuits are said to be switched. The salient characteristic of a switched virtual circuit (SVC) is that a host may set up and delete such a VC dynamically without the involvement of a network administrator. 

Note that an SVC should more accurately be called a “signalled” VC, since it is the use of signalling (not switching) that distinguishes an SVC from a PVC.

There are several things to note about virtual circuit switching:

• Since host A has to wait for the connection request to reach the far side of the network and return before it can send its first data packet, there is at least one RTT of delay before data is sent.
• While the connection request contains the full address for host B (which might be quite large, being a global identifier on the network), each data packet contains only a small identifier, which is only unique on one link. Thus, the per-packet overhead caused by the header is reduced relative to the datagram model.
• If a switch or a link in a connection fails, the connection is broken and a new one will need to be established. Also, the old one needs to be torn down to free up table storage space in the switches.
• The issue of how a switch decides which link to forward the connection request on has been similarities with the function of a routing algorithm.

For example, an X.25 network—a packet-switched network that uses the connection-oriented model—employs the following three-part strategy:

1. Buffers are allocated to each virtual circuit when the circuit is initialized.
2. The sliding window protocol is run between each pair of nodes along the virtual circuit, and this protocol is augmented with flow control to keep the sending node from overrunning the buffers allocated at the receiving node.
3. The circuit is rejected by a given node if not enough buffers are available at that node when the connection request message is processed.

Source Routing

A third approach to switching that uses neither virtual circuits nor conventional datagrams is known as source routing. The name derives from the fact that all the information about network topology that is required to switch a packet across the network is provided by the source host.

There are various ways to implement source routing. One would be to assign a number to each output of each switch and to place that number in the header of the packet. The switching function is then very simple: For each packet that arrives on an input, the switch would read the port number in the header and transmit the packet on that output. However, since there will in general be more than one switch in the path between the sending and the receiving host, the header for the packet needs to contain enough information to allow every switch in the path to determine which output the packet needs to be placed on.

Three ways to handle headers for source routing: (a) rotation; (b) stripping; (c) pointer. The labels are read right to left.