Bharat Doshi

Cooperative Service Level Agreement

With ever increasing mobility of the workforce and of the communication infrastructure itself, we will continue to see the growth in diversity of network domains and diversity of applications. The global information grid (GIG) is a prime example of such a networking environment. In most networking scenarios, the component network domains are owned and operated by different organizations. This paper discusses how to deliver end-to-end service with requisite quality of service (QoS) to such networks so that overall mission goals are maximally achieved with minimal cost while allowing autonomous operations within each component network. In particular, this paper proposes a cooperative service level agreements (SLA) approach and describes related SLA processes and roles

Sensor Network Design for Underwater Surveillance

In this paper we study a number of design aspects for a large sensor field whose mission is to protect coastal waters by detecting objects like submarines. The system is buoy-based and distributed over a littoral area. The field has a number of gateway nodes that have the capability to transmit off-field through a satellite or an airplane. The system uses active sonar with acoustic sources sending acoustic energy (pings) and a number of sensor nodes receiving energy reflected by potential targets as well as non-target objects that may be present in the field, false alarms due to this clutter is controlled through proper fusion of sensor detections. In this paper, we discuss two important issues that need to be addressed when designing such sensor network: (1) efficient in-network fusion of the sensor detections to improve the overall system-wide detection performance in terms of probability of detection and probability of false alarm. (2) effects of gateway placements to the communications performance such as the average delay and packet drop rates. Simulation results are presented to characterize the trade-offs among different design choices

A hybrid end-to-end QOS architecture for heterogeneous networks (like the global information grid)

In this paper, we discuss an architecture and several mechanisms for meeting the quality of service (QoS) requirements of a set of very diverse applications over networks of very heterogeneous network domains. Many of these domains have significant mobility and temporal variation in resource capacities. While DoD s global information grid (the GIG) is a prime example of a network of this type, commercial Internet and intranets increasingly face similar QoS requirements and operate in similar heterogeneous and dynamic environments. Meeting these challenges and still maintaining high resource efficiency, involves mechanisms in data, control, and management planes. In addition to meeting the QoS requirements of applications, these mechanisms also need to honor the importance level of the user and mission involved in the application. Thus, these mechanisms also perform policy based triage function when the requirements of all users and applications cannot be satisfied due to unexpectedly high demand or unexpected reduction in capacity due to failure, weather condition, etc. The architecture we propose here groups applications into service classes, each with its QoS requirements. Any specific instance of an application has an importance level (IL) determined by organizational policy. We use the requirements, one set for each service class-IL combination, as key drivers for designing a hybrid architecture involving session level admission control and preemption for session oriented service classes, and a packet level marking scheme allowing differential treatment of packets from different service classes and from different ILs.

Inter-domain routing with multi-dimensional QoS requirements

External border gateway protocol (eBGP) is the ubiquitous protocol used today for advertisement of reachability information and for route selection among administrative domains (autonomous systems or ASs) in the Internet. However, many emerging needs in commercial and military networking have exposed limitations of the current eBGP. In particular, these IP networks of the future carries a very diverse mix of applications, with very diverse quality of service (QoS) requirements (in the broad sense of the phrase). Some of these networks also have a very diverse set of component networks (wireless and wireline, fixed and mobile with different degrees of mobility, long lived and short term ad-hoc) and some of the component networks may be very dynamic in their service capabilities. These scenarios call for enhancing eBGP to allow for multi-topology and QoS-aware routing, using several QoS metrics in decision making. In such an enhanced BGP, more than one route (or path vector) may be propagated in BGP_UPATE messages, helping optimize with respect to different QoS metrics as needed by different traffic types. In this paper, we discuss such an enhancement of eBGP. We develop details of advertisements, route thinning, and route selection needed to make the eBGP multi-topology and QoS-aware in the sense described above. We take the eBGP and internal BGP (iBGP) advertisement and route selection process and identify any modifications needed at each stage. We also discuss interactions between eBGP and iBGP and between BGP and the interior (intra-domain) gateway protocol (IGP) needed to make the approach work end-to-end. We also discuss alternative ways to ensure that packets follow the selected end-to-end routes (both within and between domains). In particular, the potential uses of MPLS, source routing, tunneling, and DiffServ/ToS bits for this purpose are discussed in the paper

Precedence and Quality of Service (QoS) Handling in IP Packet Networks

In military networks, precedence and preemption (P&P) and quality of service (QoS) are both required. QoS refers to meeting the performance requirements of an application (packet delay, packet loss, packet delay variation, service availability, connection set up time, connection acceptance ratio, etc.) while P&P refers to meeting the QoS requirements of the highest importance applications under congestion conditions, in which there are not enough network resources to satisfy the QoS requirements of all applications. In particular, preemption refers to taking away network resources from a lower precedence application in order to give the resources to a higher precedence application. Per Hob Behaviors (PHBs) are used by network nodes to implement QoS. A service level agreement (SLA) is a QoS contract that is used to assure that the QoS requirements of admitted applications are met by the network transport for defined traffic profiles. SLAs and P&P are often perceived to have an adversarial relationship. This paper proposes a way to resolve this conflict by making QoS PHBs precedence-aware, while remaining agnostic of the specific PHB mechanism

QoS routing in multi-domain networks

With increasing number of mission critical applications and associated traffic volume on commercial and DoD IP networks, it is becoming increasingly critical for the network infrastructure to be more aware of the quality of service (in a broad sense of the word) requirements and use this awareness in routing, differential treatment of packets, in management-level resource provisioning and allocation. In this paper, we focus on the implication of these requirements on the routing architecture and protocols. In particular, we argue that the inter domain routing protocols like the border gateway protocol (BGP) and its integration with the interior gateway protocol (IGP) need to account for the diversity in the QoS requirements of the applications, users, and mission. The solution we suggest involves allowing multiple inter-AS routes for the same source destination pair, propagation of multiple routes, and of multiple QoS-affecting metrics in the BGP UPDATE messages. These enhancements allow routes to be matched with application requirements and hence allow satisfaction of applications QoS requirements at minimal cost. In order to keep the solution scalable, we limit the propagation of information to non-dominated routes, and design a small set of QoS metrics. Details of the protocols and quantitative analysis will be presented in future papers. Our approach, along with IntServe and DiffServe based QoS capabilities, and traffic engineering extensions of intra domain routing protocols, provide a framework for end-to-end QoS-aware networking.

On a global information grid simulation platform for investigations of end-to-end performance

We have embarked upon the development of a reusable global information grid (DID) simulation platform in order to support GIG architecture, design and engineering. We followed a classical systems engineering approach to the design of the GIG simulation platform. We have identified a set of useful performance studies to perform related to GIG architecture and design. We present our approach to the design of our GIG simulation platform and the development of a GIG topology model allowing for automated simulation configurations. As an example of the platforms utility, we briefly discuss our initial studies of BGP performance in GIG-like environments. We conclude with a discussion of follow on work on our GIG simulation platform.

Gateways for mobile routing in tactical network deployments

The US Department of Defense (DoD) is developing a Network Centric Warfighting (NCW) capability. Key to the deployment of NCW capabilities is the development of scalable networks supporting end user mobility. Initial network deployments operate either At-The-Halt (ATH) or On-the-Move (OTM) with preplanned movements. This is consistent with current networking capabilities with respect to large scale mobile network capabilities and protocols. However, future architectures and capabilities should allow for more flexible mobility models allowing for more flexible and robust NCW capabilities. We investigate hierarchical network models which are comprised of a high bandwidth, planned mobile core network interconnecting subtending more mobile end user networks. Standard IP routing and name and location services are assumed within the core network. The subtending and mobile end user networks rely upon a highly scalable (from a mobility perspective) Beacon-Based routing architecture. The interface between the core and subtending mobile networks relies upon network concepts being developed within the Internet Engineering Task Force (IETF), specifically from IPv4 and IPv6 mobility and the Host Identity Protocol (HIP) rendezvous service for mobile networks. We discuss the advantageous of this architecture in terms of mobility, scalability, current DoD network plans and commercial protocol development.

Design considerations for sensor networks with gateways

In a large class of sensor network deployments, a small subset of the sensors covering the sensor field is equipped with special communications capability to communicate with operators outside the sensor field. These sensors play the role of gateways for off-field communication in the sense that all communications to- or out of the field is through these nodes, and the other non-gateway nodes are only capable of sensor-to-sensor communication. This design achieves a lower cost by concentrating expensive communication devices in a small subset of nodes. An important problem in designing such gateway-based sensor networks is determining the number of gateway nodes needed, their location in the sensor field, and the automation of the sensor-to-gateway association for off-field communication. These design considerations are addressed in this paper. In determining the number of gateways the tradeoff is between performance and cost. As the number of gateways increases, less traffic load is placed on each gateway and its surrounding nodes, resulting in longer network lifetime and larger off-field aggregate transmission capacity. However, with a larger number of gateways the network may be too costly to deploy as gateway nodes are more expensive than non-gateway sensor nodes. We develop and analyze models that allow us to determine the optimal number of gateways and their location in the sensor field. We also provide initial results with respect to determining the needed number of fusion nodes. While the presence of multiple gateways offers a higher degree of off-field communication reliability, a sensor will need to select one of the gateways at a time for off-field communication. In this paper, we also propose a dynamic sensor-to-gateway association protocol. Based on current energy levels, the distributed protocol dynamically assigns sensors to gateways in such a way that the overall transmission load is balanced among the different gateway regions over the lifetime of the sensor field.

Border gateway protocol 4 (BGP4) performance over intermittent satellite links

We performed an investigation of the border gateway protocol 4 (BGP-4) (Y. Rekhter and T. Li, 1995) behavior in the presence of intermittent satellite channels. Specifically, we developed an analytic model of BGP-4 state transitions and use this model to estimate bandwidth consumption of BGP-4 over the intermittent channels. We compare our models estimates with fidelity simulation results of bandwidth consumption of BGP-4. We briefly discuss this work in the context of the global information grid (GIG) and our ongoing efforts to develop a system wide simulation model of the GIG.