Computer Science and Engineering

In the IEEE 802.11 Power Save Mode (PSM) specified for Independent Basic Service Set (IBSS), time is divided into beacon intervals. At the beginning of each beacon interval, each station in the power save mode periodically wakes up for a duration called announcement traffic indication message (ATIM) window. The stations that have successfully transmitted ATIM frame within the ATIM window will compete to transmit data frame in the rest of the beacon interval. The transmission of an ATIM frame depends on the CSMA/CA mechanism specified in the IEEE 802.11 DCF. The probability of a successful transmission of an ATIM frame has a great impact on the performance of IBSS in power save mode. This paper presents an analytical model to calculate the throughput using the success probability of an ATIM frame transmission in ATIM window of fixed size. The simulation results validate the accuracy of this analytical analysis.

Self calibration based blind attacks are quite accurate to detect JPEG domain steganography. Efficient prediction of cover image statistics from its stego version is an important requirement in calibration based attacks. Domain separation i.e. separating embedding domain from the steganalytic domain, is used as a countermeasure against such calibration based attacks because it complicates the cover image prediction process. Most of the domain separation techniques in the recent past are based on randomization of embedding location. In this paper, a new domain separation technique is proposed which is based on parametric discrete cosine transform. In the proposed scheme, image transformation is randomized to separate embedding domain from the steganalytic domain. A comprehensive set of experiments is carried out to justify its applicability and evaluate its performance against JPEG domain steganalyis.

Ad hoc networks, which have seen drastic increase in their usage scenarios and convergence of different applications' traffic lately, are getting ready to support QoS and secure traffic. Existing protocols for ad hoc networks provide little or no support for QoS and security. We present a new routing protocol 'CRESQ' for ad hoc networks, with adequate support for QoS using resource reservation. Ad hoc networks pose certain unique challenges for QoS implementation, in addition to those presented by the wireline and wireless scenarios where resource reservation, high speed of the mobile nodes, and frequent hand-offs from the base-stations are the issues. Additional challenges in ad hoc networks are attributed to mobility of intermediate nodes, absence of routing infrastructure and, low bandwidth and computational capacity of the nodes. The proposed protocol 'CRESQ' takes all these limitations into account and provides a general framework for implementation of QoS. CRESQ, which is a cluster based routing protocol, effectively uses the clustering to minimize the routing overhead and to provide QoS guarantees. CRESQ is a source routing protocol, with an ability for localized route recovery to minimize route and QoS re-establishment delay. Simulations have shown that CRESQ performs considerably better than the existing protocols, in terms of packet drop ratio and routing overhead.

A cluster-based route discovery and dynamic route management protocol for ad hoc networks with inaccurate information for given QoS (Quality of Service) requirements has been proposed and compared with existing routing protocols. The proposed scheme is much more scalable, the route discovery is faster

Ad hoc networks are characterized by multihop wireless connectivity, frequently changing network topology and the need for efficient dynamic routing protocols. Recent comparative studies between Ad hoc on demand routing protocols like Ad hoc On demand Distance Vector routing (AODV) and Dynamic Source Routing (DSR) (the two on demand routing protocols for Ad hoc networks) have shown that AODV performs better than DSR for high mobility cases but faces the problem of high routing and MAC load as compared to DSR. This is because DSR resort to aggressive use of caching of routes while AODV does not. In this paper, we have incorporated caching of routes in AODV with the aim to reduce the routing and MAC load of AODV without changing the basic structure of the protocol. A detailed simulation model with MAC and physical layer models is used to study the effect of caching of routes in AODV and to compare its performance with AODV without cache and DSR. We show that caching of routes in AODV can lead to significant reduction in routing and MAC load as well as in delay in delivering the packet as compared to AODV without much compromise in terms of packet delivery fraction.

Clustering in ad hoc networks provides significant support for implementation of QoS and security, by overcoming inherent network deficiencies (like lack of infrastructure etc.). In a clustered network architecture, the ad hoc network is divided into groups of nodes called clusters. The clusters are dynamically maintained and reconfigured using specific protocols and algorithms. In this paper, we describe a clustering technique which has been designed as a part of the CRESQ([12]) routing algorithm for ad hoc networks. CRESQ's clustering is segregated into two parts, the initial clustering and the cluster management. Initial Clustering creates the clusters during the network formation and cluster management maintains the clusters throughout the network lifetime. The main focus of this paper is the cluster management algorithm which is designed to be distributed in nature, adaptive to practicalities such as loss of packets and tolerant to mobilities during the algorithm execution. The algorithm makes no assumptions about the availability of information concerning transmission range, node mobilities etc. This paper also discusses certain performance metrics for judging the efficiency of clustering in ad hoc networks. We also present simulation results for the cluster management algorithm, along with a comparison with some existing algorithms.

In the Internet, with many competing networks each trying to optimise its own bandwidth, a stub network has limited knowledge about user demands, available network resources and routing policies of other networks. This uncertainty makes the task of interdomain traffic engineering for a stub network very challenging. The basic aim of a stub network connected to multiple ISPs (multihomed) is to load balance its traffic among its various edge links. Our goal in this work is to distribute the incoming traffic of a multihomed stub network among its various edge links. The focus is on networks that primarily download traffic from the Internet. Regulating the incoming traffic is difficult since it will require to influence the behaviour of the remote destinations. We performed a systematic analysis of our problem and showed that even a restricted instance of the problem is NPcomplete. We proposed simple, low-cost route control techniques to load balance traffic by reallocating the routes of outgoing traffic. The techniques were validated using synthetic as well as actual data collected under numerous traffic load conditions. Results show that we can achieve significant improvement in load balancing with minimum traffic re-assignments. Moreover, the proposed techniques neither require any third party assistance nor changes to existing protocols and network setup. This makes our schemes easily deployable in real networks.

A model for storage and retrieval of relevant documents during a document production workflow in an ofJice is presented. The objective of the workflow is to review a document by multiple reviewers, where each reviewer contributes her own comments during review based on other documents containing standing rules, precedents etc, which is termed, the context. Storage of such documents is modeled as a Page Cube and query languages analogous to relational languages have been proposed for easy and efficient retrieval.

In order to bridge the speed disparity at the LAN-WAN interface, enterprises install a number of policy-based devices to enforce administrative policies over the WAN access. The conventional techniques of bandwidth management concentrate only on the outgoing traffic from the local network, since that is the only traffic that can be controlled. However it is usually the incoming traffic that causes bottlenecks. This paper proposes a simple load balancing approach to optimize the use of WAN links in a multihomed environment. Incoming traffic on a link is controlled by diverting outgoing traffic from some nodes to other links. Incoming bandwidth is measured, thresholds are defined, and traffic diversion takes place to maintain the thresholds. We present a scheme to implement this. A prototype implementation has been done on an educational campus network. Preliminary experiments have shown that there are significant improvements in performance. These results are presented. Proposed future work to extend the scheme to an ISP network is described

Identifying the sources of a Distributed Denial-of-Service (DDoS) attack is among the hardest problems in the Internet security area. Use of reflectors in a DDoS attack makes the problem particularly difficult as the actual sources of attack (slaves) are camouflaged. In this paper we propose a modification to the traceback of Multiple Hash DERM [13], an effective deterrent against DDoS attacks. The proposed modification -DERM Traceback Module, helps the victim not only in perpacket filtering of the attack traffic but also in identifying the nature of the attack and tracking the actual slaves involved in reflector attacks and mixed DDoS attacks. As in [13], the marking procedure at routers is simple, bandwidth overhead is nil, processing requirements at the victim are minimal, and a limited number of packets are required to carry out the traceback procedure.

A consistent backup, preserving data integrity across files in a file system, is of utmost importance for the purpose of correctness and minimizing system downtime during the process of data recovery. With the present day demand for continuous access to data, backup has to be taken of an active file system, putting the consistency of the backup copy at risk.

Denial-of-Service (DoS) and Distributed Denial-of-Service (DDoS) attacks are becoming increasingly sophisticated with few practical solutions available. In this paper we consider the issue of filtering reflector based DoS attacks and of identifying attackers. For reflector based attacks, a Signature Conflict Triggered Filtering (SCTF) scheme based on Deterministic Edge Router Marking (DERM) was proposed. We suggest an enhancement to make the 3-way handshake in SCTF stateless and call it Fast-SCTF. We then propose a framework using BGP for a single-packet handshake. We demonstrate that our proposed scheme is space efficient, more secure, robust and it requires very little cooperation among autonomous systems.

A number of multi-version file systems have been proposed and implemented in recent years. Besides presenting an option for the user to recover from mistakes, versioning is also useful for crash recovery. This paper presents a multiversion file system. The design is based on providing versioning support for selected files in a system, and in an environment where such multi-version files are long lasting and change relatively slowly. As with most multi-version systems, the aim is to ensure that access to the current version of a file are not affected by the presence of multiple versions. Writes are made "in-place" with the older data being shifted to another location. The main situation identified is an office environment where multiple versions of documents for rules, regulations, etc., are common. Separate strategies are proposed for maintaining version information of attributes and of the addresses of blocks of different versions. An analysis of the performance of the system using trace data from University of California, Berkeley, shows that the degradation in performance of accessing the current version in the presence of multiple versions of a file, is small and acceptable. The schemes presented allow older versions to be accessed with an overhead that is less as compared to other multiversion systems whose information is available in the literature.

Load balancing traffic of a multihomed stub network among its various egress links can significantly improve performance by allowing the network to route its traffic around congested paths. In, several centralized route control schemes to load balance traffic were proposed. It was shown that employing these route control schemes resulted in significant improvement in round-trip times. We carry this exploration further by developing distributed algorithms to load balance traffic. We show theoretically that the expected performance of the distributed version is the same as the centralized routing scheme. We further validate the schemes using both synthetic as well as actual traffic traces.

Devices which have limited computing resources but are capable of networking have become increasingly important constituents of distributed environments. In order to maximize the advantages of their networking capability, it is important to provide efficient methods to access modify and share data. In this paper we position distributed filesystems as a possible solution to this problem. We discuss the constraints imposed by such environments on traditional distributed filesystem design parameters. We describe the design and implementation of a distributed filesystem (lean-dfs) for such environments.

Bursty, high-bandwidth, real-time and mission-critical applications are driving the need for better differentiation of service in the campus network beyond that provided by traditional best-effort services. The ability to provide these services depends on the availability of cohesive management tools that can be used to provision and monitor sets of routers in a coordinated manner. To facilitate the development of such a management tool it is essential to first define a framework that aims at setting up generic QoS-enabling components across the campus network. In this paper we study such a scheme for enabling QoS in a campus network and discuss the design and implementation of a management tool that eliminates the need for device-by-device command streams. We also investigate the performance improvement gained by a proposed change in the three-tier mechanism.