Packet-switching communication network and method of design

Claim 1

A method for designing a packet-switching communication network having at least two nodes and at least one link from node-to-node for transmitting packets, the method comprising the steps of: (1) setting an initial capacity for each link of said network; (2) determining a most congested link based on delay distribution functions, each delay distribution function defining a relationship between a delay time and a probability that packets reach a destination node from a source node within said delay time; (3) modifying the capacity of said most congested link; and (4) repeating steps 2 and 3 until the value of each distribution function at a predetermined maximum allowable delay time is equal to or greater than a selected probability value for all paths from source nodes to destination nodes, resulting in a network which allows all packets to reach a destination node from a source node within said predetermined maximum allowable delay time with said selected probability value or greater probability value.

Claim 2

A method for designing a packet-switching communication network having at least two nodes and at least one link from node-to-node for transmitting packets, the method comprising the steps of: (1) setting an initial capacity for each link of said network; (2) determining a most congested link based on delay distribution functions, each delay distribution function defining a relationship between a delay time and a probability that packets reach a destination node from a source node within said delay time; (3) modifying the capacity of said most congested link; and (4) repeating steps 2 and 3 until the value of each distribution function at a predetermined maximum allowable delay time is equal to or greater than a selected probability value for all paths from source nodes to destination nodes, resulting in a network which allows all packets to reach a destination node from a source node within said predetermined maximum allowable delay time with said selected probability value or greater probability value.

Claim 6

A method for assigning capacities to a plurality of links of a network which has a plurality of paths between nodes, each of said plurality of paths including at least one of said plurality of links, the method comprising the steps of: (1) setting an initial capacity for each of said links of the network; (2) finding a minimum probability path among all of the paths of the network, said minimum probability path having a minimum probability P that packets are transmitted from a beginning to an end of said minimum probability path within a predetermined maximum allowable delay time; (3) finding a minimum probability link among all of the links included in said minimum probability path, said minimum probability link having a minimum probability that packets are transmitted from a beginning to an end of said minimum probability link within said predetermined maximum allowable delay time; (4) modifying the capacity of said minimum probability link; and (5) repeating steps 3 and 4 if said minimum probability P of said minimum probability path is less than a selected probability value, resulting in a design of a network which allows all packets to reach a destination node from a source node within said predetermined maximum allowable delay time with said selected probability value or greater probability value.

Claim 7

A method for assigning capacities to a plurality of links of a network which has a plurality of paths between nodes, each of said plurality of paths including at least one of said plurality of links, the method comprising the steps of: (1) setting an initial capacity for each of said links of the network; (2) finding a minimum probability path among all of the paths of the network, said minimum probability path having a minimum probability P that packets are transmitted from a beginning to an end of said minimum probability path within a predetermined maximum allowable delay time; (3) finding a minimum probability link among all of the links included in said minimum probability path, said minimum probability link having a minimum probability that packets are transmitted from a beginning to an end of said minimum probability link within said predetermined maximum allowable delay time; (4) modifying the capacity of said minimum probability link; and (5) repeating steps 3 and 4 if said minimum probability P of said minimum probability path is less than a selected probability value, resulting in a design of a network which allows all packets to reach a destination node from a source node within said predetermined maximum allowable delay time with said selected probability value or greater probability value.

Claim 11

A method of designing a packet-switching communication network having at least two nodes and at least one link from node-to-node for transmitting packets, the method comprising the steps of: (1) setting an initial capacity for each link of the network; (2) calculating a value of Fx pq;k (t k ) at t k =T k for all p, q, k where Fx pq;k (t k ) is a distribution function of random variable X pq;k which defines a relationship between a delay time t k and a probability that packets of a traffic class k reach a destination node q from a source node p within said delay time t k and T K is a predetermined maximum allowable delay time for each traffic class k; (3) determining an end-to-end path defined by a node pair p O -q O of the of the network for transmission of packets of a traffic class k o , which satisfies the equation; Fx.sub.p0q0;k0 (T.sub.k0)=min {Fx.sub.pq;k (T.sub.k)} for all p, q, k, where min {Fx.sub.pq;k (T.sub.k)} represents the minimum value of {Fx.sub.pq;k (T.sub.k)}; (4) calculating a value of Fx i;k0 at t k =T k0 for all links l i included in said end-to-end path, where Fx i;k0 (t k ) is a distribution function which defines relationship between a delay time t k and a probability that packets of a traffic class k 0 are transmitted from a beginning to an end of a link l i within said delay time t k and T kO is a predetermined maximum allowable delay time for a traffic class k 0 ; (5) determining a link l i0 which satisfies the equation: Fx.sub.i0;k0 (T.sub.k0)=min {Fx.sub.i;k0 (T.sub.k0)} for all links l.sub.i included in said end-to-end path, where min {Fx.sub.i;k0 (T.sub.k0)} represents the minimum value of {Fx.sub.i;k0 (T.sub.k0)}; (6) increasing the capacity of said link l i0 ; and (7) repeating steps (2) through (6) if Fx p0q0;k0 (T k0 )<P min , where P min is fixed probability value, resulting in a design of a network which allows all packets to reach a destination node from a source node within said predetermined maximum allowable delay time for all traffic classes with said fixed probability value or greater probability value.

Claim 12

A method of designing a packet-switching communication network having at least two nodes and at least one link from node-to-node for transmitting packets, the method comprising the steps of: (1) setting an initial capacity for each link of the network; (2) calculating a value of Fx pq;k (t k ) at t k =T k for all p, q, k where Fx pq;k (t k ) is a distribution function of random variable X pq;k which defines a relationship between a delay time t k and a probability that packets of a traffic class k reach a destination node q from a source node p within said delay time t k and T K is a predetermined maximum allowable delay time for each traffic class k; (3) determining an end-to-end path defined by a node pair p O -q O of the of the network for transmission of packets of a traffic class k o , which satisfies the equation; Fx.sub.p0q0;k0 (T.sub.k0)=min {Fx.sub.pq;k (T.sub.k)} for all p, q, k, where min {Fx.sub.pq;k (T.sub.k)} represents the minimum value of {Fx.sub.pq;k (T.sub.k)}; (4) calculating a value of Fx i;k0 at t k =T k0 for all links l i included in said end-to-end path, where Fx i;k0 (t k ) is a distribution function which defines relationship between a delay time t k and a probability that packets of a traffic class k 0 are transmitted from a beginning to an end of a link l i within said delay time t k and T kO is a predetermined maximum allowable delay time for a traffic class k 0 ; (5) determining a link l i0 which satisfies the equation: Fx.sub.i0;k0 (T.sub.k0)=min {Fx.sub.i;k0 (T.sub.k0)} for all links l.sub.i included in said end-to-end path, where min {Fx.sub.i;k0 (T.sub.k0)} represents the minimum value of {Fx.sub.i;k0 (T.sub.k0)}; (6) increasing the capacity of said link l i0 ; and (7) repeating steps (2) through (6) if Fx p0q0;k0 (T k0 )<P min , where P min is fixed probability value, resulting in a design of a network which allows all packets to reach a destination node from a source node within said predetermined maximum allowable delay time for all traffic classes with said fixed probability value or greater probability value.

Claim 16

An apparatus for designing a packet-switching communication network having at least two nodes and at least one link from node-to-node for transmitting packets, the apparatus comprising: (1) means for setting an initial capacity for each link of said network; (2) means for determining a most congested link based on delay distribution functions, each delay distribution function defining a relationship between a delay time and a probability that packets reach a destination node from a source node within said delay time; (3) means for modifying the capacity of said most congested link; and (4) means for repeatedly applying said means for determining and means for modifying such that the value of each distribution function at a predetermined maximum allowable delay time is equal to or greater than a selected probability value for all paths from source nodes to destination nodes, resulting in a network which allows all packets to reach a destination node from a source node within said predetermined maximum allowable delay time with said selected probability value or greater probability value.

Claim 17

An apparatus for designing a packet-switching communication network having at least two nodes and at least one link from node-to-node for transmitting packets, the apparatus comprising: (1) means for setting an initial capacity for each link of said network; (2) means for determining a most congested link based on delay distribution functions, each delay distribution function defining a relationship between a delay time and a probability that packets reach a destination node from a source node within said delay time; (3) means for modifying the capacity of said most congested link; and (4) means for repeatedly applying said means for determining and means for modifying such that the value of each distribution function at a predetermined maximum allowable delay time is equal to or greater than a selected probability value for all paths from source nodes to destination nodes, resulting in a network which allows all packets to reach a destination node from a source node within said predetermined maximum allowable delay time with said selected probability value or greater probability value.

Claim 18

An apparatus for assigning capacities to a plurality of links of a network which has a plurality of paths between nodes, each of said plurality of paths including at least one of said plurality of links, the apparatus comprising: (1) means for setting an initial capacity for each of said links of the network; (2) means for finding a minimum probability path among all of paths of the network, said minimum probability path having a minimum probability P that packets are transmitted from a beginning to an end of said minimum probability path within a predetermined maximum allowable delay time; (3) means for finding a minimum probability link among all of the links included in said minimum probability path, said minimum probability link having a minimum probability that packets are transmitted from a beginning to an end of said minimum probability link within said predetermined maximum allowable delay time; (4) means for modifying the capacity of said minimum probability link; and (5) means for repeatedly applying said means for finding a minimum probability link and said means for modifying if said minimum probability P of said minimum probability path is less than a selected probability value, resulting in design of a network which allows all packets to reach a destination node form a source node within said predetermined maximum allowable delay time with said selected probability value or greater probability value.

Claim 19

An apparatus for assigning capacities to a plurality of links of a network which has a plurality of paths between nodes, each of said plurality of paths including at least one of said plurality of links, the apparatus comprising: (1) means for setting an initial capacity for each of said links of the network; (2) means for finding a minimum probability path among all of paths of the network, said minimum probability path having a minimum probability P that packets are transmitted from a beginning to an end of said minimum probability path within a predetermined maximum allowable delay time; (3) means for finding a minimum probability link among all of the links included in said minimum probability path, said minimum probability link having a minimum probability that packets are transmitted from a beginning to an end of said minimum probability link within said predetermined maximum allowable delay time; (4) means for modifying the capacity of said minimum probability link; and (5) means for repeatedly applying said means for finding a minimum probability link and said means for modifying if said minimum probability P of said minimum probability path is less than a selected probability value, resulting in design of a network which allows all packets to reach a destination node form a source node within said predetermined maximum allowable delay time with said selected probability value or greater probability value.

Claim 20

An apparatus for designing a packet-switching communication network having at least two nodes and at least one link from node-to-node for transmitting packets, the apparatus comprising: (1) means for setting an initial capacity for each link of the network; (2) means for calculating a value of Fx pq;k (t k ) at t k =T k for all p, q, k, where FX pq;k (t k ) is a distribution function of random variable X pq;k (t k ) which defines a relationship between a delay time t k and a probability that packets of traffic class k reach a destination node q from a source node p within said delay time t k and T k is a predetermined maximum allowable delay time for each traffic class k; (3) means for determining an end-to-end path defined by a node pair p 0 -q 0 of the network for transmission of packets of a traffic class k 0 , which satisfies the equation: Fx.sub.p0q0;k0 (T.sub.k0)=min {Fx.sub.pq;k (T.sub.k)} for all p, q, k, where min {Fx.sub.pq;k (T.sub.k)} represents the minimum value of {Fx.sub.pq;k (T.sub.k)}; (4) means for calculating a value of Fx i;k0 (T k ) at t k =T k0 for all links l i included in said end-to-end path, where Fx i;k0 (t k ) is a distribution function which defines a relationship between a delay time t k and a probability that packets of a traffic class k 0 are transmitted from a beginning to an end of a link l i within said delay time t k and T k0 is a predetermined maximum allowable delay time for a traffic class k 0 ; (5) means for determining a link l i0 which satisfies the equation: Fx.sub.i0;k0 (T.sub.k0)=min {Fx.sub.i;k0)} for all links l.sub.i included in said end-to-end path, where min {Fx.sub.i;k0 (T.sub.k0)} represents the minimum value of {Fx.sub.i;k0 (T.sub.k0)} (6) means for increasing the capacity of said link l i0 ; and (7) means for repeatedly applying said means 2 through 6 if Fx p0q0;k0 (T k0 )<P min , where P min is a fixed probability value, resulting in a design of a network which allows all packets to reach a destination node from a source node within said predetermined maximum allowable delay time for all traffic classes with said fixed probability value or greater probability value.

Claim 21

An apparatus for designing a packet-switching communication network having at least two nodes and at least one link from node-to-node for transmitting packets, the apparatus comprising: (1) means for setting an initial capacity for each link of the network; (2) means for calculating a value of Fx pq;k (t k ) at t k =T k for all p, q, k, where FX pq;k (t k ) is a distribution function of random variable X pq;k (t k ) which defines a relationship between a delay time t k and a probability that packets of traffic class k reach a destination node q from a source node p within said delay time t k and T k is a predetermined maximum allowable delay time for each traffic class k; (3) means for determining an end-to-end path defined by a node pair p 0 -q 0 of the network for transmission of packets of a traffic class k 0 , which satisfies the equation: Fx.sub.p0q0;k0 (T.sub.k0)=min {Fx.sub.pq;k (T.sub.k)} for all p, q, k, where min {Fx.sub.pq;k (T.sub.k)} represents the minimum value of {Fx.sub.pq;k (T.sub.k)}; (4) means for calculating a value of Fx i;k0 (T k ) at t k =T k0 for all links l i included in said end-to-end path, where Fx i;k0 (t k ) is a distribution function which defines a relationship between a delay time t k and a probability that packets of a traffic class k 0 are transmitted from a beginning to an end of a link l i within said delay time t k and T k0 is a predetermined maximum allowable delay time for a traffic class k 0 ; (5) means for determining a link l i0 which satisfies the equation: Fx.sub.i0;k0 (T.sub.k0)=min {Fx.sub.i;k0)} for all links l.sub.i included in said end-to-end path, where min {Fx.sub.i;k0 (T.sub.k0)} represents the minimum value of {Fx.sub.i;k0 (T.sub.k0)} (6) means for increasing the capacity of said link l i0 ; and (7) means for repeatedly applying said means 2 through 6 if Fx p0q0;k0 (T k0 )<P min , where P min is a fixed probability value, resulting in a design of a network which allows all packets to reach a destination node from a source node within said predetermined maximum allowable delay time for all traffic classes with said fixed probability value or greater probability value.

Claim 22

A method for constructing a packet-switching communication network having at least two nodes and at least one link from node-to-node for transmitting packets, the method comprising the steps of: (1) setting an initial capacity for each link of said network; (2) determining a most congested link based on delay distribution functions, each delay distribution function defining a relationship between a delay time and a probability that packets reach a destination node from a source node within said delay time; (3) modifying the capacity of said most congested link; (4) repeating steps (2) and (3) until the value of each distribution function at a predetermined maximum allowable delay time is equal to or greater than a selected probability value for all paths from source nodes to destination nodes; and (5) establishing said link between nodes having capacity based on the foregoing steps, resulting in a network which allows all packets to reach a destination node from a source node within said predetermined maximum allowable delay time with said selected probability value or greater probability value.

Claim 23

A method for constructing a packet-switching communication network having at least two nodes and at least one link from node-to-node for transmitting packets, the method comprising the steps of: (1) setting an initial capacity for each link of said network; (2) determining a most congested link based on delay distribution functions, each delay distribution function defining a relationship between a delay time and a probability that packets reach a destination node from a source node within said delay time; (3) modifying the capacity of said most congested link; (4) repeating steps (2) and (3) until the value of each distribution function at a predetermined maximum allowable delay time is equal to or greater than a selected probability value for all paths from source nodes to destination nodes; and (5) establishing said link between nodes having capacity based on the foregoing steps, resulting in a network which allows all packets to reach a destination node from a source node within said predetermined maximum allowable delay time with said selected probability value or greater probability value.

Claim 25

A method for constructing a plurality of links in a network, which has a plurality of paths between nodes, each of said plurality of paths including at least one of said links, comprising the steps of: (1) setting an initial capacity for each of said links of the network; (2) finding a minimum probability path among all of the paths of the network, said minimum probability path having a minimum probability P that packets are transmitted from a beginning to an end of said minimum probability path within a predetermined maximum allowable delay time; (3) finding a minimum probability link among all of the links included in said minimum probability path, said minimum probability link having a minimum probability that packets are transmitted from a beginning to and end of said minimum probability link within said predetermined maximum allowable delay time; (4) modifying the capacity of said minimum probability link; (5) repeating steps (3) and (4) if said minimum probability P of said minimum probability path is less than a selected probability value; and (6) establishing said links between nodes having capacities based on the foregoing steps, resulting in a design of a network which allows all packets to reach a destination node from a source node within said predetermined maximum allowable delay time with said selected probability value or greater probability value.

Claim 26

A method for constructing a plurality of links in a network, which has a plurality of paths between nodes, each of said plurality of paths including at least one of said links, comprising the steps of: (1) setting an initial capacity for each of said links of the network; (2) finding a minimum probability path among all of the paths of the network, said minimum probability path having a minimum probability P that packets are transmitted from a beginning to an end of said minimum probability path within a predetermined maximum allowable delay time; (3) finding a minimum probability link among all of the links included in said minimum probability path, said minimum probability link having a minimum probability that packets are transmitted from a beginning to and end of said minimum probability link within said predetermined maximum allowable delay time; (4) modifying the capacity of said minimum probability link; (5) repeating steps (3) and (4) if said minimum probability P of said minimum probability path is less than a selected probability value; and (6) establishing said links between nodes having capacities based on the foregoing steps, resulting in a design of a network which allows all packets to reach a destination node from a source node within said predetermined maximum allowable delay time with said selected probability value or greater probability value.

Claim 28

A method of constructing a packet-switching communication network having at least two nodes and at least one link from node-to-node for transmitting packets, the method comprising the steps of: (1) setting an initial capacity for each link of the network; (2) calculating a value of Fx pq;k (t k ) at t k =T k for all p, q, k, where Fx pq;k (t k ) is a distribution function of random variable X pq;k which defines a relationship between a delay time t k and a probability that packets of a traffic class k reach a destination node q from a source node p within said delay time t k and T k is a predetermined maximum allowable delay time for each traffic class k; (3) determining an end-to-end path defined by a node pair p 0 -q 0 of the network for transmission of packets of a traffic class kO, which satisfies the equation: Fx.sub.p0q0;k0 (T.sub.k0)=min {Fx.sub.pq;k (T.sub.k)} for all p, q, k, where min {Fx.sub.pq;k (T.sub.k)} represents the minimum value of {Fx.sub.pq;k (T.sub.k)}; (4) calculating a value of Fx i;k0 (t k ) at t k =T k0 for all links l i included in said end-to-end path, where Fx i;k0 (t k ) is a distribution function which defines a relationship between a delay time t k and a probability that packets of a traffic class k 0 are transmitted from a beginning to an end of a link l i within said delay time t k and T k0 is a predetermined maximum allowable delay time for a traffic class k 0 ; (5) determining a link l i0 which satisfies the equation: Fx.sub.i0;k0 (T.sub.k0)=min {Fx.sub.i;k0 (T.sub.k0)} for all links l.sub.i included in said end-to-end path, where min {Fx.sub.i;k0 (T.sub.k0)} represents the minimum of {Fx.sub.i;k0 (T.sub.k0)}; (6) increasing the capacity of said link l i0 ; and (7) repeating steps (2) through (6) if Fx p0q0;k0 (T k0 )<P min , where P min is fixed probability value; and (8) establishing said links between nodes having capacities based on the foregoing steps, resulting in a design of a network which allows all packets to reach a destination node from a source node within said predetermined maximum allowable delay time for all traffic classes with said fixed probability value or greater probability value.

Claim 29

A method of constructing a packet-switching communication network having at least two nodes and at least one link from node-to-node for transmitting packets, the method comprising the steps of: (1) setting an initial capacity for each link of the network; (2) calculating a value of Fx pq;k (t k ) at t k =T k for all p, q, k, where Fx pq;k (t k ) is a distribution function of random variable X pq;k which defines a relationship between a delay time t k and a probability that packets of a traffic class k reach a destination node q from a source node p within said delay time t k and T k is a predetermined maximum allowable delay time for each traffic class k; (3) determining an end-to-end path defined by a node pair p 0 -q 0 of the network for transmission of packets of a traffic class kO, which satisfies the equation: Fx.sub.p0q0;k0 (T.sub.k0)=min {Fx.sub.pq;k (T.sub.k)} for all p, q, k, where min {Fx.sub.pq;k (T.sub.k)} represents the minimum value of {Fx.sub.pq;k (T.sub.k)}; (4) calculating a value of Fx i;k0 (t k ) at t k =T k0 for all links l i included in said end-to-end path, where Fx i;k0 (t k ) is a distribution function which defines a relationship between a delay time t k and a probability that packets of a traffic class k 0 are transmitted from a beginning to an end of a link l i within said delay time t k and T k0 is a predetermined maximum allowable delay time for a traffic class k 0 ; (5) determining a link l i0 which satisfies the equation: Fx.sub.i0;k0 (T.sub.k0)=min {Fx.sub.i;k0 (T.sub.k0)} for all links l.sub.i included in said end-to-end path, where min {Fx.sub.i;k0 (T.sub.k0)} represents the minimum of {Fx.sub.i;k0 (T.sub.k0)}; (6) increasing the capacity of said link l i0 ; and (7) repeating steps (2) through (6) if Fx p0q0;k0 (T k0 )<P min , where P min is fixed probability value; and (8) establishing said links between nodes having capacities based on the foregoing steps, resulting in a design of a network which allows all packets to reach a destination node from a source node within said predetermined maximum allowable delay time for all traffic classes with said fixed probability value or greater probability value.

Claim 31

A method for designing a packet-switching communication network having at least two nodes and at least one link from node-to-node for transmitting packets of a plurality of traffic classes, the method comprising the steps of: (1) setting an initial capacity for each link of said network; (2) obtaining, for each link, a capacity increment ΔC d i required for satisfying a delay distribution constraint; (3) obtaining, for each link, a capacity increment ΔC l i required for satisfying a loss probability constraint; (4) finding a largest one of said increments ΔC d i and said increments ΔC l i , and modifying the capacity of a link associated with said largest increment; and (5 ) repeating steps (2) and (3) until a relationship of ΔC d i =ΔC l i =0 is satisfied for all the links, resulting in a network which satisfies both said delay distribution constraint and said loss probability constraint.

Claim 32

A method for designing a packet-switching communication network having at least two nodes and at least one link from node-to-node for transmitting packets of a plurality of traffic classes, the method comprising the steps of: (1) setting an initial capacity for each link of said network; (2) obtaining, for each link, a capacity increment ΔC d i required for satisfying a delay distribution constraint; (3) obtaining, for each link, a capacity increment ΔC l i required for satisfying a loss probability constraint; (4) finding a largest one of said increments ΔC d i and said increments ΔC l i , and modifying the capacity of a link associated with said largest increment; and (5 ) repeating steps (2) and (3) until a relationship of ΔC d i =ΔC l i =0 is satisfied for all the links, resulting in a network which satisfies both said delay distribution constraint and said loss probability constraint.

Claim 38

An apparatus for designing a packet-switching communication network having at least two nodes and at least one link from-node-to-node for transmitting packets of a plurality of traffic classes, the apparatus comprising: (1) means for setting an initial capacity for each link of said network; (2) means for obtaining, for each link, a capacity increment ΔC d i required for satisfying a delay distribution constraint; (3) means for obtaining, for each link, a capacity increment ΔC l i required for satisfying a loss probability constraint; (4) means for finding a largest one of said increments ΔC d i and said increments ΔC l i , and modifying the capacity of a link associated with said largest increment; and (5) means for repeatly applying said means for obtaining ΔC d i and ΔC l i until a relationship of ΔC d i =ΔC l i =0 is satisfied for all the links, resulting in a network which satisfies both said delay distribution constraint and said loss probability constraint.

Claim 39

An apparatus for designing a packet-switching communication network having at least two nodes and at least one link from-node-to-node for transmitting packets of a plurality of traffic classes, the apparatus comprising: (1) means for setting an initial capacity for each link of said network; (2) means for obtaining, for each link, a capacity increment ΔC d i required for satisfying a delay distribution constraint; (3) means for obtaining, for each link, a capacity increment ΔC l i required for satisfying a loss probability constraint; (4) means for finding a largest one of said increments ΔC d i and said increments ΔC l i , and modifying the capacity of a link associated with said largest increment; and (5) means for repeatly applying said means for obtaining ΔC d i and ΔC l i until a relationship of ΔC d i =ΔC l i =0 is satisfied for all the links, resulting in a network which satisfies both said delay distribution constraint and said loss probability constraint.

Claim 40

A method of constructing a packet-switching communication network having at least two nodes and at least one link from-node-to-node for transmitting packets of a plurality of traffic classes, the method comprising the steps of: (1) setting an initial capacity for each link of said network; (2) obtaining, for each link, a capacity increment ΔC d i required for satisfying a delay distribution constraint; (3) obtaining, for each link, a capacity increment ΔC l i required for satisfying a loss probability constraint; (4) finding a largest one of said increments ΔC d i and said increments ΔC l i , and modifying the capacity of a link associated with said largest increment; (5 ) repeating steps (2) and (3) until a relationship of ΔC d i =ΔC l i =0 is satisfied for all the links; and (6) establishing said link having capacity based on the foregoing steps in said network, resulting in a network which satisfies both said delay distribution constraint and said loss probability constraint.

Claim 41

A method of constructing a packet-switching communication network having at least two nodes and at least one link from-node-to-node for transmitting packets of a plurality of traffic classes, the method comprising the steps of: (1) setting an initial capacity for each link of said network; (2) obtaining, for each link, a capacity increment ΔC d i required for satisfying a delay distribution constraint; (3) obtaining, for each link, a capacity increment ΔC l i required for satisfying a loss probability constraint; (4) finding a largest one of said increments ΔC d i and said increments ΔC l i , and modifying the capacity of a link associated with said largest increment; (5 ) repeating steps (2) and (3) until a relationship of ΔC d i =ΔC l i =0 is satisfied for all the links; and (6) establishing said link having capacity based on the foregoing steps in said network, resulting in a network which satisfies both said delay distribution constraint and said loss probability constraint.