Link capacity sharing for throughput-blocking optimality

An electronically implemented method for sharing link capacity among elastic traffic connections in a network, comprising the steps of:

• receiving network traffic input parameters;
• determining a maximum number of admissible elastic traffic connections on said link based on said network traffic input parameters, based on maximizing the throughput of in-progress elastic traffic connections with respect to the number of admissible elastic traffic connections under at least one blocking probability constraint for requested elastic traffic connections, said input parameters including at least one throughput/blocking constraint imposed on said elastic traffic;
• receiving requests for elastic traffic connections on a link in said network;
• and exercising admission control for said requested elastic traffic connections based on said determined maximum number of admissible elastic traffic connections.

A device for sharing link capacity in a network, comprising:

• means for receiving network traffic input parameters;
• means for dividing said link capacity into a first part common to elastic traffic and non-elastic traffic and a second part dedicated for elastic traffic, based on at least part of said network traffic input parameters;
• and
• means determining at least one admission control parameter for said elastic traffic based on said division of link capacity and at least part of said network traffic input parameters.

An electronic system for allocating link bandwidth among and within different traffic classes in a network, at least one of said different traffic classes being an elastic traffic class, said electronic system comprising:

• a processor configured for:
• partitioning said link bandwidth into a first part common to all traffic classes, and a second part dedicated to connections of said elastic traffic class(es);
• and
• allocating said second dedicated part of said link bandwidth to connections of said elastic traffic class(es) based on link bandwidth utilization under at least one throughput/blocking constraint imposed on the connections of said elastic traffic class(es).

An electronic system for sharing link capacity among elastic traffic connections in a network, comprising a processor responsive to network traffic input parameters and configured for:

• determining a maximum number of admissible elastic traffic connections based on at least one throughput/blocking constraint imposed on said elastic traffic, based on minimizing the blocking probability of requested elastic traffic connections with respect to said maximum number of admissible elastic traffic connections under at least one throughput-threshold constraint for in-progress elastic traffic connections;
• receiving request for elastic traffic connections;
• and
• exercising admission control for said:requested elastic traffic connections based on said determined maximum number of admissible elastic traffic flows.

A method for link bandwidth sharing in an admission-control enabled IP network, comprising the step of:

• applying a call-level model of a link carrying a number n, where n is an integer equal to or greater than 1, of elastic traffic classes for dimensioning the link bandwidth sharing for throughput-blocking optimality, wherein said call-level model is defined by:
• said link having a predetermined bandwidth capacity C;
• and
• for each one of said n elastic traffic classes, the elastic traffic being modeled as:
• i) having a peak bandwidth requirement and a minimum bandwidth requirement;
• ii) occupying the maximum possible bandwidth within said peak and minimum bandwidth requirements;
• and
• iii) being associated with at least one of a minimum accepted throughput and a maximum accepted blocking probability.

A method for determining a steady state distribution of Markov chain describing the dynamics of a network link carrying traffic of a number of traffic classes including a non-adaptive elastic traffic class, said method comprising the steps of:

• determining a link capacity sharing policy for said link;
• determining a multi-dimensional Markov chain having a set of feasible states for the number of active connections of said traffic classes according to constraints imposed by said link sharing policy;
• calculating an initial approximation of the steady state distribution of said Markov chain based on Markov chain product form calculations;
• and
• iteratively determining the steady state distribution starting from said initial approximation of the steady state distribution.