Indo-German Joint project titled “Next Generation Secure Technology for Globally Distributed Enterprises”
Routing assumes a complex dimension in the absence of hierarchy in the network. The limits of the complexity reached in the case of peer-to–peer communications. However, the same can be turned to an advantage for the application we have in mind. Taking the specific example of the Business Process Outsourcing industry, we aim to improve its performance and security. This requires a combined investigation of P2P systems, routing and security issues. The study of routing mechanisms, particularly overlay techniques that superimpose fast and secure pathways on the basis of traffic characterizations, is another important subject matter of our joint project.
Principal Investigators:
Dr. M. Sethumadhavan ( Amrita, India) and
Prof. Anja Feldmann (TUB, Germany) .
Design and analysis of cryptographically secure pseudorandom number Generators
(Sponsored by Defence Research and Development Organisation)
A cryptographically secure pseudorandom number generator is a pseudorandom number generator with properties that makes it suitable for use in cryptography. CSPRNGs are designed explicitly to resist determined mathematical reverse engineering. In this project, we are studying the designs of PRNGs based on number theory. We concentrate in fast generation of the following generators: Power generators and subset sum generators.
Research Scholar:
Ms. C. Kavitha
Development of Fast and Effective Cryptographic Techniques (Completed)
(Sponsored by Indian Space Research Organization)
A major issue in the security analysis of stream ciphers is the quality assessment of key streams. In other words, we need to know how close the key stream is to true randomness. Key stream guaranteeing an adequate security level must meet various requirements such as possessing good statistical randomness properties and a high complexity (in a suitable sense), so that the key stream cannot be inferred from a small portion of its terms. In this project we concentrate on linear complexity and related complexity measures for key sequences.
Analysis of Word based Stream Ciphers
(Sponsored by Advanced Data Processing and Research Institute (ISRO))
For the assessment of the quality of multisequence for word based stream ciphers, we need to know the behaviour of joint linear complexity profile of random multisequence over a finite field. The joint linear complexity and joint linear complexity profile of multisequences have received a lot of attention recently. In this project we propose to study further the properties of multisequences in the same line. This will be of direct relevance to the analysis of the stream cipher candidates of the eSTREAM cipher project.
Research Scholars:
Ms. Sindhu
Mr. M. Chungath Srinivasan
Fast implementation of finite field Arithmetic for Elliptic curve Cryptography
(Sponsored by National Research Technical Organization)
In this project we attempt to address various VLSI design related issues with a special focus on low area and low power implementation of the finite filed arithmetic. We intend to select a few algorithms with the potential for application in the area. Their realization with FGPA will be studied, analyzed and compared. The focus will be on the identification of algorithms and parameter combinations which will deliver optimum performance.
Research Scholar:
Ms. Kodeswari
Student Projects
Hardware Implementation of Montgomery multiplication
Security requirements in present day communications are achieved through efficient algorithms such as RSA encryption. The fundamental operation of such algorithms is modular exponentiation which involves repeated modular multiplications. However, the high bit-length operations, required to provide adequate security, challenges the design of an efficient hardware. In this Project we study the
x^2 mod
n generator and its hardware implementation in an FPGA through Montgomery Modular Multiplication.
Gaurav Kumar
Dual Watermarking On Still Images
Digital watermarking is the technique in which a visible/invisible signal (watermark) is embedded in a multimedia document for copyright protection. In this case we embed a watermark image on the original image. We propose a watermarking scheme called “dual watermarking”. Dual watermark is a combination of a visible watermark and an invisible watermark. We present an approach for still image watermarking in which the watermark embedding process incorporates a model of the human visual system (HVS) and at the same time it should be robust to various possible attacks. An invisible image is embedded in the original image in order to prove ownership in the worst case where the visible watermark is totally tampered/removed. The original unmarked image is required to extract the watermark.
Manimegalai. P
Shalini. N
Varshinigayathri. T
Mahalakshmi. V. R
Network Intrusion Detection and Prevention System
Today’s computer networks are becoming more and more complex. The types of data communicated over the network are of various types. More and more business today are operating over the networks to connect to their offices, business partners, customers etc, across the globe. The Value of the data, privacy of information and network security are of prime magnitude consideration of network implementations today. The traditional security methods like firewall, NAT etc are becoming insufficient for protecting networks today
Dynamic Security solutions are needed to protect networks today from Hackers and unwanted access. Intrusion Detection / Prevention is one of such most recent dynamic technology to secure networks which handle vital data.
This project is aimed to develop IDS/IPS system to work with Cisco Routers and Switches to detect intrusive activity and block such connectivity based on some signatures.
Binit Singh
Bhuvaneshwari. S
Archana and
Kamala
Scheme for combining pollution in Peer-to-Peer Systems
Peer-to-Peer systems have seen a rapid increase in popularity over the past few years, and consume a large portion of the internet band width. These systems, mainly used for file -sharing, are generating a lot of buzz in the research community as there are many security issues that are coming to the fore.
Pollution in P2P systems
Invalid or bad chunks of data transferred under an otherwise valid description, is termed as pollution. Pollution of content on these networks is a major cause of concern as it not only causes inconvenience to users, but also wastes a large fraction of the bandwidth.
There are many trust based schemes which have been developed for combating pollution. Based on these schemes, a peer downloads content from another peer, only if it trusts that peer. We propose to develop a new scheme which combines ideas from these existing models ( eg. XRep protocol) along with a measure of nearness of the peer (which will be provided by the ISP hosted Oracle System).We base this model on the assumption that the closer a peer is topologically, the more it can be trusted. We will be implementing this scheme on top of Gnutella, which is a popular P2P system. We then plan to simulate this implementation on SSFNet (Scalable Simulation Framework) and analyze its performance with respect to the reduction in pollution.
Vignesh R
Deployment and Experimentation of Nearest Neighbor Discovery Scheme (The Oracle) for distributed systems in PlanetLab
Peer-to-Peer networks and protocols have gained immense popularity on the Internet in recent times.P2P traffic forms more than 50% of the Internet traffic. In a P2Pnetwork the users of the system known as the nodes connect directly with each other at the application layer. Most of the applications/protocols available for P2P operations like Gnutella, BitTorrent and eDonkey etc. choose their neighbors in a random fashion. As such, two nodes which appear to be neighbors at the Overlay (Peer-to-Peer) level may actually be very far away from each other at the underlay (ISP) level. Therefore there arises a need for the localization of the P2P traffic at the underlay level as well so as to minimize the inter ISP traffic, to reduce costs as well as improve the service quality.
This project is about implementing the Nearest Neighbor discovery scheme known as the “Oracle” which is a central server hosted at a well known IP address within an ISP. This will be implemented in the Internet wide test bed, and deployment platform known as PlanetLab. The project aims at deploying and testing this proposed scheme for the most popular file sharing (P2P) protocol on the Internet known as Bit Torrent. The overall scheme for the project will work as follows:
A central Oracle server will be deployed at a publicly known IP Address.
- Instances of BitTorrent applications normal (unbiased) as well as the ones which work in conjunction with the Oracle Service, the biased ones will be deployed across the PlanetLab nodes in a phased manner.
- For the biased case the BitTorrent program that has been deployed will make use of the Oracle for discovery of peers which are in the neighborhood of the given node to achieve localization of the P2P traffic.
- Then the performance of the two cases (biased as well as unbiased) would be compared based on predefined parameters and the results obtained from the PlanetLab deployment.
This will enable us to analyze the performance improvement of BitTorrent nodes and traffic localization in ISPs on an Internet wide scale.
Binit Singh Bhatia
A Reliable and Robust nearest Neighbor Selection Scheme for Ad hoc Networks Using Zone Routing Concept.
The continuing proliferation of traffic across today’s networks demands the deployment of an efficient and robust means of data transmission. Ad Hoc Network nodes accept responsibility for managing the network and routing of data. Routing of data, the most intensive task, is performed in a distributed manner. To send data to a destination who is not an immediate neighbor, a node must first find a route to that destination. Intermediate nodes co-operate to forward packets from the source to the destination. Ad hoc networks consist of devices which are free to enter and exit at any time and move at any rate. This has implications for the performance of the network most notably with regard to stability of routes and delivery of packets.
Our ambition is to utilize the Zone Routing Protocol concept from ZRP and adapt it to find nearest neighbor, which proactively maintains routing information for a local neighborhood (routing zone), while reactively acquiring routes to destinations beyond the routing zone.
We would also try to compare/correlate the zone concept with the centralized server concept –Oracle, which the ISPs offer to find the nearest neighbor.
Although various protocols exists which elaborate on techniques based on nearest neighbor concept most of them have not considered the reliability and robustness issues which is important in peer-to-peer systems.
This work looks up at the effectiveness of the zone concept to find the nearest neighbor in the most reliable and robust manner.
Sreejith K.S
Implementation of an ISP-hosted nearest neighbor selection Web Server.
