scispace - formally typeset
Search or ask a question

Showing papers by "Somnath Ghosh published in 2007"


Journal ArticleDOI
TL;DR: It is concluded that administration of U at an acute nephrotoxic dose caused oxidative stress in brain and bone manifested as lipid peroxidation and histopathological damage.
Abstract: Although the kidneys are the main target organs for uranium (U) toxicity, recent studies have shown that U can cross the blood-brain barrier to accumulate in the brain. Uranyl nitrate (U-238)induced oxidative damage was investigated in brain and bone of Wistar rats after intraperitoneal injection of uranyl nitrate at acute doses either nephrotoxic (576 microg of U/kg body weight) or subnephrotoxic (144 microg U/kg body weight). The health effects of U administration at 576 microg of U/kg body weight were seen in terms of decrease in food intake and no gain in body weight compared to respective controls. These alterations were correlated with increased lipid peroxidation as measured by thiobarbituric acid reactive substances in rat brain and bone. However, at lower dosage of U (144 microg U/kg body weight), no significant lipid peroxidation was observed in brain and bone. Histological examination of U-treated (576 microg of U/kg body weight) rat brain tissues showed marked and diffuse cystic degeneration and a similar pattern in histological alterations was observed in kidneys in treated animals; whereas no significant histological change was observed in rat brains and kidney treated with a lower dose of U (144 microg U/kg body weight). It is concluded that administration of U at an acute nephrotoxic dose caused oxidative stress in brain and bone manifested as lipid peroxidation and histopathological damage.

19 citations


Journal ArticleDOI
TL;DR: In this article, the authors exploit a simple and accurate matrix method to analyze the effects of introducing a linear chirp either in thickness or in refractive index of the cladding layers on the propagation characteristics of 1D photonic band gap planar Bragg reflection waveguides (BRWs).
Abstract: We exploit a simple and accurate matrix method to analyze the effects of introducing a linear chirp either in thickness or in refractive index of the cladding layers on the propagation characteristics (loss and dispersion) of 1D photonic band gap planar Bragg reflection waveguides (BRWs). We show that an appropriate chirp in the otherwise periodic claddings of finite extent BRWs could be gainfully exploited to tailor its leakage loss and waveguide dispersion features. In particular, we theoretically demonstrate that for some reported sample BRWs, leakage loss and waveguide dispersion could be significantly reduced by a factor of 30–50 and by about two orders of magnitude, respectively as compared to un-chirped BRWs. Furthermore, we also show that in contrast to un-chirped BRWs, how chirped BRWs could be designed with attractive feature like much less number of cladding layers and nearly wavelength independent losses. Our analysis and proposal should serve as a useful design tool to tailor the propagation characteristics of BRWs.

12 citations


Proceedings ArticleDOI
14 May 2007
TL;DR: The paper proposes a modification of the protocol to eliminate the need for some nodes in S-MAC to stay awake longer than the other nodes, and improves the energy efficiency and increases the life span of a wireless sensor network.
Abstract: S-MAC is a robust medium access control protocol for wireless sensor networks which is designed to be energy efficient. S-MAC nodes reduce energy expenditure by periodically turning off their receivers (sleeping) in a coordinated manner. Nodes form virtual clusters based on common sleep schedules. However, some nodes in S-MAC may have to wake up more often than the other nodes. This paper demonstrates in some wireless sensor networks using S-MAC, a significant proportion of the nodes may have to stay awake much longer than envisaged. The paper then proposes a modification of the protocol to eliminate the need for some nodes to stay awake longer than the other nodes. The modified version improves the energy efficiency and increases the life span of a wireless sensor network.

6 citations


Proceedings ArticleDOI
11 Jul 2007
TL;DR: A performance evaluation of the hypertext transfer engines using the simulated network environment ns, a high- bandwidth symmetric network and a dial-up connection and an analysis of the HTTP-MPLEX response encoding scheme comparing the metrics of response time, header retrieval time and total object transfer time is given.
Abstract: HTTP-MPLEX is a header compression and response encoding scheme for HTTP. It is designed to compress traditional HTTP requests and multiplex any number of responses to a single stream, thus improving application use of TCP, reducing the number of competing parallel connections and improving the time required to retrieve responses from a server. HTTP 1.1 is limited to transmitting one response per connection, HTTP-MPLEX enables many responses to be transmitted at the same time via a single TCP connection. To analyse the effect of HTTP-MPLEX request compression and focus on the effectiveness of response multiplexing in a symmetric network environment, a client and server was written in C++ with plug-in versions of HTTP 1.1 and HTTP-MPLEX engines. This paper is a performance evaluation of our hypertext transfer engines using the simulated network environment ns (2.29-snapshot-20050921), a high- bandwidth symmetric network and a dial-up connection. An analysis of the HTTP-MPLEX response encoding scheme comparing the metrics of response time, header retrieval time and total object transfer time is given. Snap-shots of www.cnn.com, www.whitehouse.gov, www.latrobe.edu.au and a photo gallery of forty eight tiles are used in experimentation as sample web pages to retrieve using the hypertext transfer engines.

4 citations


Proceedings ArticleDOI
14 May 2007
TL;DR: A performance evaluation of the hypertext transfer engines using both the simulated network environment ns (2.29-snapshot- 20050921) and an Asymmetric Digital Subscriber Line (ADSL) connection to the Internet.
Abstract: HTTP-MPLEX [1] is a header compression and response encoding scheme for HTTP. It is intended to speed up response time for multiple simultaneous HTTP transactions and improve application layer use of TCP by reducing the number of parallel connections and sustaining response bursts. HTTP- MPLEX achieves these objectives by reducing request size and prioritizing responses during multiplexing. To evaluate the performance of HTTP-MPLEX relative to HTTP 1.1 [2] in an asymmetric network environment; we developed a client and server in C++ with plug-in versions of our HTTP 1.1 and HTTP-MPLEX engines. We present in this paper a performance evaluation of our hypertext transfer engines using both the simulated network environment ns (2.29-snapshot- 20050921) and an Asymmetric Digital Subscriber Line (ADSL) connection to the Internet. We used snap-shots of www.cnn.com, www.whitehouse.gov, www.latrobe.edu.au and a photo gallery of forty eight tiles as sample web pages to retrieve using our hypertext transfer engines.

2 citations


Journal ArticleDOI
TL;DR: In this article, the a-C:H buffer layer was employed to improve the crystalline property of ZnO thin film for the membrane film bulk acoustic resonator (FBAR).

1 citations


Proceedings ArticleDOI
14 May 2007
TL;DR: It is shown that AODV-S is not immune to DoS attack using modified hop count, and a solution to this problem using an ordering algorithm is provided.
Abstract: Protecting the network layer is an important problem in an ad hoc network. Several base network later protocols were built under the assumption that nodes forming an ad hoc network do not have enough power or processing capabilities to launch a powerful malicious attack. Thus they do not offer any immunity. AODV-S offers enhancement to the AODV protocol by incorporating protection against several routing attacks. However in this paper, we show that AODV-S is not immune to DoS attack using modified hop count. We also provide a solution to this problem using an ordering algorithm.

1 citations