Unintrusive communication of status in a packet network in heavy traffic

We consider a packet switched network in the situation where communication resources are used close to capacity. Such heavy traffic may seem to present a dilemma. On one hand, at each node, the usefulness of status information about queues at other nodes is manifest. On the other hand, since the limitation of transmission resources causes backup, heavy load seems to be the worst situation in which to expend still more communication resources to convey status information. Under extremely general assumptions on inter-arrivals and services, a scaling appropriate for queueing processes in networks under heavy traffic has been established. Under these assumptions, we demonstrate that the status of the entire network can be communicated throughout the network, perfectly, in real time, without influencing the scaled queueing process. So, within a precise mathematical setting, we see that there is no dilemma: status can be conveyed at a negligible cost in a network operating at heavy load. Most of today's computer networks are designed for light-to-moderate loading. Yet heavy traffic analysis is growing in relevance, as is explained. A brief introduction to the subject of convergence of queueing systems is included.

[1]  D. Iglehart,et al.  Multiple channel queues in heavy traffic. I , 1970, Advances in Applied Probability.

[2]  G. J. Foschini,et al.  A Basic Dynamic Routing Problem and Diffusion , 1978, IEEE Trans. Commun..

[3]  John P. Lehoczky,et al.  Gaussian Approximations to Service Problems: A Communication System Example. , 1976 .

[4]  Martin I. Reiman,et al.  Open Queueing Networks in Heavy Traffic , 1984, Math. Oper. Res..

[5]  K. Parthasarathy PROBABILITY MEASURES IN A METRIC SPACE , 1967 .

[6]  Yu. V. Prokhorov Convergence of Random Processes and Limit Theorems in Probability Theory , 1956 .

[7]  Gerard J. Foschini Equilibria for diffusion models of pairs of communicating computers - Symmetric case , 1982, IEEE Trans. Inf. Theory.

[8]  F. Knight Essentials of Brownian Motion and Diffusion , 1981 .

[9]  Ward Whitt,et al.  Heavy-Traffic Limits for Queues with Many Exponential Servers , 1981, Oper. Res..

[10]  G. J. Foschini,et al.  Sharing Memory Optimally , 1983, IEEE Trans. Commun..

[11]  J. Harrison,et al.  Reflected Brownian Motion on an Orthant , 1981 .

[12]  Hisashi Kobayashi,et al.  Modeling and analysis , 1978 .

[13]  K. Mani Chandy,et al.  Approximate Methods for Analyzing Queueing Network Models of Computing Systems , 1978, CSUR.

[14]  Hisashi Kobayashi,et al.  Queueing Models for Computer Communications System Analysis , 1977, IEEE Trans. Commun..

[15]  A. Borovkov Some Limit Theorems in the Theory of Mass Service , 1964 .

[16]  Patrick Billingsley,et al.  Weak convergence of measures - applications in probability , 1971, CBMS-NSF regional conference series in applied mathematics.

[17]  Eugene Wong,et al.  Stochastic processes in information and dynamical systems , 1979 .

[18]  D. Heyman An approximation for the busy period of the M/G/1 queue using a diffusion model , 1974, Journal of Applied Probability.

[19]  J.K. Wolf,et al.  Computer-communication network design and analysis , 1979, Proceedings of the IEEE.