Vous êtes sur la page 1sur 12

PNNI ROUTING HIERARCHY

PNNI specifies a hierarchical structure for maintaining

topological information. Reduces overheads of flat structure. Provides an efficient mechanism for routing. Flat Structure:
Each node has to have the knowledge of the complete

topology of network, Maintains reachablility information for every other node in the network. Huge routing tables needs to be maintained. Management of relatively small sized networks is easier, but increased size may lead to overheads.
In hierarchical network, each node maintains a partial

view(extent of network topology visible to a node) of

In PNNI routing hierarchy, sets of nodes are grouped into

peer groups. Nodes in a peer group share information with each other and have complete information about the topology of the peer group. At the lowest level of hierarchy, There are many peer groups. At the next level, each peer group is represented by a Logical Group Node (LGN). Collection of LGNs again form higher-level peer groups and this is a recursive procedure. Lowest-Level Node:
At the bottom of the PNNI hierarchy is the lowest-level node,

or simply, node. In simple terms, a node can be viewed as an ATM switch used to support PNNI protocol.

The lowest-level nodes are identified(i.e. addressed) using a

22-byte address structure. This address structure consists of three parts:


The first byte represents level indicator and specifies the level of

the peer group to which the node belongs. Second byte is set to a value of 160. This value is chosen to distinguish between a node address and a LGN address (different value then this). The remaining 20 bytes represent the ATM end-system address
(AESA). o These AESA addresses are modelled after the NSAP address format. o For the purpose of routing, only first 19 bytes of AESA are used. o 20th byte (SEL) has a local significance only and is ignored by PNNI routing.

Logical Node:
Refers to either a lowest-level node or to a Logical Group

Logical Group Node (LGN):


An abstract representation of a peer group as a single point

for the purpose of representing that peer group at a particular level (but not at the lowest level) in the PNNI routing hierarchy LGN (N.1) hides the internal details of its child peer group (PG(1)), but still provides adequate information to LGNs (N.2, N.3, N.4, etc) of the same peer group to reach every node that is abstracted by the LGN.
(This summarized view is adequate for all nodes of N.1 to reach any

node of say N.2)

An LGN, that resides at a particular level represents a group

of interconnected LGNs of lower levels. This provides a recursive definition. An LGN is uniquely identified by a logical group node 14bytes 6 bytes identifier. 1 byte 1byte Level Peer Group ID End-System ID Zero Both lowest-level node address and LGN address are referred to as NODE IDENTIFIERs. (Depending upon

Peer Group (PG):


A set of logical nodes which are grouped for the purpose of

creating a routing hierarchy. Concept of peer group is applicable at any level of the hirerachy. Identified using Address prefix field(APF). APF can have a maximum length of 13 bytes. APF alongwith 1-byte level indicator field, forms PEERGROUP IDENTIFIER (PGID). Lower the value of level indicator, higher will be the level of the peer group.
Peer Group Leader (PGL):
A node of a peer group that performs extra group of

collecting, aggregating and building data that is suitable to represent entire peer group as a single node in a higher-level peer group. In a peer group, there can be atmost 1 PGL.

The election is based upon nodes leadership priority (highest

priority). In case of ties, node with higher node identifier is selected. PGL election is a continuous process.
Links:
All links are called Logical links.
A logical link represents the connection between two logical

nodes. Includes Physical links and Virtual path connections(VPC). Within a peer group is horizontal link. Links connecting nodes of two different peer groups is called an outside link. Nodes connected via an outside link are called border nodes.

PNNI TOPOLOGICAL INFORMATION


Each node in the PNNI hierarchy maintains topological

information, which is used by nodes to perform routing functionality. Topological information is categorized as:
Nodal Information State Information Reachability Information

Nodal Information:
Encompasses general information about a node. Elements are: Nodes AESA: Refers to ATM address required for reaching this address. Leadership priority: Desirability of node to become PGL. Preferred PGL node ID: Refers to node ID of that node of which this node wants to be a PGL

State Information:
Maintained to determine whether a new connection must be

accepted or not (checks QoS). State Information parameters falls in two categories:
Metric

Attribute

Attributes:
Considered individually. Non-Additive link parameters. E.g. Available cell rate.

Metric:
Requires that the values of state parameters of all the links

and nodes along a given path be combined before a decision on the combined value is made. Additive parameters. E.g. Maximum Cell Transfer Delay (maxCTD)

Reachability Information:

Identifies the node reachable through a PNNI routing

domain. Topological Information that binds reachable addresses to nodes within the PNNI routing domain. Depending upon whether PNNO protocol is used for routing exchange or not, reachablility information is classified as :
Interior information
Exterior information Interior information:
Local knowledge of reachability within the PNNI domain. Contains information like Port ID, Address Information Length,

Address Information Count, etc.

Exterior information: Information derived from protocol exchanges outside PNNI domain. Contains information like Port ID, Address Information Length, Address

ROUTING INFORMATION FLOWS AND ROUTE COMPUTATION


How

various PNNI elements exchange information. The key procedures in PNNI network include
Node Initialization, Database Synchronization, Flooding, Information Flow Up the Hierarchy, Information Flow Down the Hierarchy,

routing

Node Initialization: As soon as, a node comes up (becomes active), it exchanges state information with its immediate neighbors. State information is carried in a Hello packet and contains, among other things, the ATM end-system address, node ID and peer group ID(PGID-the most important component). A node upon receiving a Hello packet from its neighbor , compares the received PGID value with its own PGID. If two values match, the node tries to synchronize itd topology database with neighbor who had originated the message.

Database Synchronization: Once neighboring nodes realize that they belong to the same peer group, they proceed to synchronize its topology database in order to attain identical topology database. Topology database synchronization process between two nodes can be explained as follows:
Nodes firstly exchange the PNNI Topology State Element (PTSE) header

information. (PTSE is a collection of PNNI information that is flooded among all logical nodes within a peer group) When a node receives a PTSE header information, it requests for the advertised PTSE. (PTSE header is a 20 byte field and contains elements like PTSE identifier, sequence number, checksum, remaining lifetime, etc. ) Using the elements, The PTSE header identifies a particular PTSE. After receiving the requested PTSE, the node updates its topology database with the received PTSE.

Flooding: After synchronizing the topology databases, a node in the peer group proceeds to propagate the topology information to all members of the peer

Continuous process. Flooding process: The PTSE obtained after database synchronization are bundled within a PNNI Topology State Packet (PTSP) and are transmitted to all the neighboring nodes. For each PTSE received in a PTSP packet, the receiving node acknowledges the receipt by sending acknowledgement packet. If received PTSE is more recent, then it replaces the previously maintained. PTSEs are sent periodically according to acknowledgement.

Information Flow Up the Hierarchy: Till now, information exchange was at lowest level. This discusses exchange in hierarchy. Peer group leader(PGL) has all information about its peers PG(X.Y) in peer group(PG). PGL then summarizes the topology information of PG and floods the information to logical nodes of peer group of upper level. PG(X)
Information Flow Down the Hierarchy:
Downward information can be viewed as set of address prefixes,

which each lowest-level node must know in order to determine whether a destination address is reachable in the connection

Vous aimerez peut-être aussi