In recent years, utility-oriented resource allocation for wired and wireless networks has been extensively studied. A key goal of utility based analysis is to provide evaluation criteria for efficient network operation based on subjective user assessments such as usefulness and value of data. Yet the vast majority of the prior work has focused on topics like mathematical functions (concave or otherwise) of network metrics such as bandwidth, delay, packet loss, etc., or information entropy, or user-perceived quality (e.g., MOS) for interpreting utility. In this work1, we propose an alternative mission-oriented definition and metric for utility that is based on the accuracy and speed at which tasks are completed, which we believe is closer to the intent of utility-based analysis. Like previous work, our definition supports the design and engineering of networks by mapping utility metrics to the typical network design metrics (bandwidth, delay, loss, etc). Unlike previous work, it also permits simple solutions to important questions such as the joint utility or usefulness of different data streams, the impact on the utility of one data stream by another, and cross-sensory utility (e.g., the impact of a side audio channel on an image processing task, or a side video channel on an audio processing task). We present a novel experimental approach to the design of such experiments, and provide measurement results. The results quantify the effects of information encoding and the impairments incurred during transmission through imperfect networks on the information's usefulness to end-users in terms of being able to complete tasks correctly and on time.

Due to its bandwidth efficiency, multicast makes a group-centric communication more viable in wireless ad hoc networks with limited radio resources. PIM-SM, a de facto standard multicast protocol known for its high scalability, is a good fit for a large-scale ad hoc network. However, it does not provide a robust multicast communication under RP outage and host mobility. In this paper, we propose a robust way of configuring PIM-SM using Anycast RP in wireless ad hoc networks. We analyze the impact of cardinality and locations of anycast RPs on the network performance under node mobility. Based on these observations, we find metrics for near-optimal cardinality of anycast RPs and propose a novel RP selection scheme. The proposed scheme is proven to make PIM-SM robust against mobility while satisfying QoS requirements and maintaining the scalability of PIM-SM.

To build ad-hoc networks with hundreds or even thousands of nodes, the network must be split into relatively independent layer 3 clusters. There are currently no general approaches or methodologies for the creation of domains that take into account the following important engineering constraints: balanced domains, minimal inter-domain traffic, robust network design. Moreover, there is also no contribution that specifically addresses the problem of area formation for OSPF-based ad hoc networks. By assuming that the creation of layer 3 clusters is done after layer 2 topology management has set local radio parameters, we have re-formulated the problem at hand in a graph theoretic framework and found a strong connection between the problem of area design and a classical problem in graph theory, the graph partitioning (GP) problem. The limitation of GP algorithms (GPAs) is that they are OSPF-agnostic and, therefore, are not capable of solving the problem of robust backbone network (Area 0) design. We here propose a two-step approach for OSPF area design: the first step consists of an efficient GPA that will create areas (including a rough Area 0 design), and the second step consists of ad-hoc heuristics that are able to ameliorate the Area 0 design as well as factoring in OSPF specific metrics.

The existing naming schemes for pooling resources are suitable only for static networks. We propose dynamic survivable resource pooling (DSRP) that provides a survivable naming scheme to the pool users in mobile ad-hoc networks. DSRP dynamically places the name servers (NSs) on a virtual backbone (VB): a highly distributed, scalable, and survivable mesh network formed and maintained through one-hop beacons. In this paper, DSRP focuses on pooling resources such as servers and services that are common in robotics and battlefield applications. It provides an abstraction of all the functionally equivalent servers, whereby the client can access these servers as a single entity, termed server pool. A distributed backbone mesh of NSs is used to propagate service registrations, requests, and replies.

The current SCTP specification is not aware of QoS provided by a network. As a result, it is unable to support preferential treatment of the individual streams. We introduce the concept of dividing an association into subflows (SFs) so that the modified SCTP can support different levels of QoS. We define the necessary modifications to the current SCTP for this capability: namely, new data and SACK chunk formats to implement congestion control for each SF. To avoid the problem of false sharing, we modify SCTP such that each SF has its own congestion control mechanism. Using analytic models, we show that the modified SCTP is expected to perform better than the original SCTP. We verify these results through simulation experiments, where we observed that modified SCTP is able to take advantage of the diff-serv network and provides a better throughput.

Reliable server pooling (RSP) allows a pool of redundant information sources to be viewed as a single transport endpoint and, therefore, it is able to provide persistent connections and balanced traffic for different applications. The Internet Engineering Task Force RSerPool Working Group has proposed an architecture to implement the RSP; it defines an overlay network providing an upper layer protocol or an application with a range of reliability services, from simple server selection to a fully automatic session-failover capability. The simulation experiments conducted in both wired and wireless environments show that the current version of the RSerPool works well in fixed and relatively reliable environments, but its performance worsens rapidly as the networks become more unreliable or mobile. The issues we have identified in wireless mobile ad hoc networks include network partition, high signaling overhead, difficulty in synchronization among name servers, and excessive aggressiveness in handling failures. Alternative design options for the RSP in wireless and mobile environments are introduced and evaluated.

The reliable server pooling is a framework to handle session failures and increase the system's availability by providing several reliability services: from simple server selection to a full session-failover capability. In the IETF RSerPool, servers with equivalent functionality are pooled together; when a particular server becomes unavailable, a client application can transparently switched over to another server in the pool. Extensive simulation experiments identified two main shortcomings of the RSerPool in wireless mobile networks: the inaccuracy of the failure-detection mechanism for pooled servers, and high overhead of finding name servers. We introduce and evaluate alternative mechanisms that tackle these problems.

The reliable server pooling allows redundant information sources to be viewed as a single transport endpoint, and therefore is able to provide persistent connections and balanced traffic. The IETF RSerPool Working Group has proposed an architecture to implement the reliable server pooling. We conducted a number of simulation experiments with the current definitions of the RSerPool protocols to quantify their performance in both wired and wireless environments. The simulation results show that the RSerPool works well in fixed and relatively reliable environments, but its performance worsens rapidly as the networks become more unreliable or mobile. The issues we identified in wireless mobile ad-hoc networks include network partitioning, high signaling overhead, and excessive aggressiveness in handling failures. These problems are partly due to the heavy reliance of the RSerPool architecture on the reliability of the underlying network,, which is unlikely to be guaranteed in a wireless mobile ad-hoc environment.

We report on a network traffic warehousing project at Telcordia. The warehouse supports a variety of applications that require access to Internet traffic data. The applications include Service Level Agreement (SLA), web traffic analysis, network capacity engineering and planning, and billing. We describe the design of the warehouse and the issues encountered in building the warehouse.

Materialized views (MV) at the data warehouse (DW) can be kept up to date in response to changes in data sources without accessing data sources for additional information. This process is usually refered to as “self maintenance of views”. A number of algorithms have been proposed for self maintenance of views, which use auxiliary views (AV) to keep some additional information in DW. In this paper we propose an algorithm for self maintainability of multiple MVs using the above approach. Our algorithm generates a simple maintenance query to incrementally maintain an MV along with its AV at DW. The algorithm maintains these views by minimizing the number and the size of the AVs. Our approach provides better insight into view maintenance issues by exploiting the dependencies and constraints that might exist in the data sources and multiple MVs at DW.