Body Area Networks are the networks of wireless medical sensors, deployed on a person for enabling pervasive, individualized real time health management. As BAN deals with personal health data, securing them especially their communication over the wireless link is very crucial if there is adequate security feature for the patient in the body area sensor network then the adversaries can change the actual data which will lead to wrong diagnostics and treatment of the patient in order to provide a personalized health care system. The Body Area Network along with the group key is established for the security concern where they will provide a separate key to each of the sensors that are of deployed in the patient body when this key matches with that of the health care server system the key establishment of the network

The aim of thesis is to implement the embedded sandbox techniques with32 bit RAM processor. Nowadays, many security methods are depends on a effective method in which each issues of attack is countered by a custom-made approach to eliminate the incident. Conventionally access control models are depend on the paradigm of restricting the functions of users that makes protecting users from each other or protecting system resources from users. Heterogeneous network causes some difficulties to maintain, so sandbox method is used to reduce the delay and bandwidth usage. In the proposed method, the memory allocation is required with some shape and approaches all the applications in sequence manner to store. This minimizes the delay in deleting and updating the response data"s and makes the system more efficiently. This paper describes solution for these problems through path based approach and sandboxing provides security environment to the data( or information) on Heterogeneous network services.

Abstract-In the present world internet and web search engines have become an important part in one's day-today life. For a user query, more than few thousand web pages are retrieved but most of them are irrelevant. A major problem in search engine is that the user queries are usually short and ambiguous, and they are not sufficient to satisfy the precise user needs. Also listing more number of results according to user make them worry about searching the desired results and it takes large amount of time to search from the huge list of results. To overcome all the problems, an effective approach is developed by capturing the users' click through and bookmarking data to provide personalized query recommendation. For retrieving the results, Google API is used. Experimental results show that the proposed method is providing better query recommendation results than the existing query suggestion methods.