Bandwidth contracting and risk management in communication network services

We develop economic frameworks and network architectures for bandwidth contracting and managing risks in communication network services. Majority of the traffic generated from users today is carried by wireless networks (cellular, Wi-Fi) at the edges and the Internet at the core. Therefore, we consider service provisioning in both the Internet core as well as wireless edge networks. We first address the risk management problem in end-to-end service provisioning over the Internet. Towards this end, we utilize a contract-switching Internet abstraction, in which the ISPs expose themselves to other ISPs as a set of edge-to-edge contracts. Initially, we study the efficiency of inter-domain traffic engineering using the contract-switching paradigm. Next, we consider the Internet Service Provider's (ISP's) problem of providing end-to-end (e2e) services with bandwidth guarantees. We consider the risks arising from uncertain end-to-end user demand, the possibility of route failures and contract violations, and the uncertainty in the costs involved due to other participating ISPs. We develop path-vector based end-to-end contracting solutions that maximizes the ISP's profit subject to constraints on the risk of profit. In the path-vector based approach, an ISP uses its edge-to-edge (g2g) single-domain contracts and vector of contracts purchased from neighboring ISPs as the building blocks to construct, or participate in constructing, an end-toend "contract path". We develop a spot-pricing framework for the e2e bandwidth guaranteed services utilizing this path contracting strategy, by formulating it as a stochastic optimization problem with the objective of maximizing expected profit subject to risk constraints. We present time-invariant path contracting strategies that offer high expected profit at low risks, and can be implemented in a fully distributed manner. Simulation analysis is employed to evaluate the contracting and pricing framework under different network and market conditions. Next, we propose a secondary spectrum market that allows wireless service providers to purchase spectrum access rights from another provider in the form of spot spectrum licenses of short-duration as well as derivative contracts on spot spectrum. We first address the spectrum portfolio optimization (SPO) question in this context by considering two basic types of spectrum contracts: primary contract and secondary contract. While a primary contract on a channel provides guaranteed access to the channel bandwidth (possibly at a higher per-unit price), the bandwidth available to use from a secondary contract (possibly at a discounted price) is typically uncertain/stochastic. The key problem for the buyer (service provider) in this market is to determine the amount of primary and secondary contract units needed to satisfy its uncertain user demand. We formulate the problem as one of minimizing the cost of the spectrum portfolio subject to constraints on bandwidth shortage. Two different forms of bandwidth shortage constraints are considered, namely, the demand satisfaction rate constraint, and the demand satisfaction probability constraint. While the SPO problem under demand satisfaction rate constraint is shown to be convex for all density functions, the SPO problem under demand satisfaction probability constraint is not convex in general. We derive some sufficient conditions for convexity in this case. We also discuss application of the Bernstein approximation technique to approximate a non-convex demand satisfaction probability constraint by a convex constraint. We perform a thorough simulation-based study of the single-region and the multiple-region problems for different choices of the problem parameters, and provide key insights regarding the portfolio composition. Finally, we propose a model for the market price of spot spectrum licenses, in which the uncertainty is driven by a fractional Brownian motion process. Using the stochastic calculus developed for fractional Brownian motion models, we obtain a partial differential equation governing the value of derivative contracts in the spectrum market. The derivative price is expressed as the expected value of the payoff under a risk-neutral Brownian motion dynamics. We propose a variety of derivative contracts for mitigating the risks in the market and provided insights on the value of derivative contracts for different choices of model parameters.