Saikrishnan B.

Goal of the project was to build highly available, consistent and scalable distributed system.

Highly Available:
• Multiple servers running same application and serving the clients, preventing downtime
• Leader election initiation when an existing leader goes down with randomized election time-out periods for each node

Consistent:
• Implemented Data replication to provide required consistency and to avoid single point of failure.
• Elected leader was responsible for… Show more

Goal of the project was to build highly available, consistent and scalable distributed system.

Highly Available:
• Multiple servers running same application and serving the clients, preventing downtime
• Leader election initiation when an existing leader goes down with randomized election time-out periods for each node

Consistent:
• Implemented Data replication to provide required consistency and to avoid single point of failure.
• Elected leader was responsible for replication of file chunks on majority of servers and sendin data to other clusters connected in the network
• Heartbeat between leader and followers was instrumental to detect leader failure and start new election

Scalable:
• Use of common standard i.e. standardized Protobuf messages enabled the system to add any number of nodes or even clusters
• Used DNS Servers to direct to help the client find the server which avoided unnecessary network traffic, alternate approach - polling the servers directly.
- Implemented Work sharing and Work stealing to ensure uniform distributed of work(file download request) across all nodes in the cluster.
- Implemented LRU caching for faster file downloads using Google Guava Library.
- Multithreaded Java application with server implemented using Netty (Java NIO) and Protobuff for messaging.
- Persistent storage of image chunks was done on MongoDB. Show less

Working with the Office of Emergency Services (OES), City of San Jose for developing an analytics framework for disaster risk identification and damage mitigation using Apache Spark and MLib

• Market place where user can buy and sell things with features like shopping cart for adding/check-out products and maintenance of User transactions in user history
• Performance improvement done on the HTTP server using DB Connection pooling
• Technologies used: MySQL, Node.js, Express, AngularJS

- Selected to represent San Jose State University in the CSU Hackathon conducted by GE Digita
- Trained by GE Digital to work on their product, Predix Cloud Foundry and to develop a Cloud application hosted in and making use of the Predix Cloud Services
- A Raspberry Pi connected with moisture sensors in grass, collects the moisture level data and sends the sensor data to the Predix Cloud application
- The data is then stored in the Cloud Foundry and Multivariant data analysis is… Show more

- Selected to represent San Jose State University in the CSU Hackathon conducted by GE Digita
- Trained by GE Digital to work on their product, Predix Cloud Foundry and to develop a Cloud application hosted in and making use of the Predix Cloud Services
- A Raspberry Pi connected with moisture sensors in grass, collects the moisture level data and sends the sensor data to the Predix Cloud application
- The data is then stored in the Cloud Foundry and Multivariant data analysis is performed on the data like comparison of data with the weather API data to aid in the watering decision making process for the end users
- Won the “Most Innovative Idea” award Show less

A RESTful server side component built using Node.js monitors the system health and client load. Based on the data, the server will suggest the best configuration settings like throttling, latency, number-of-retries etc for the clients. An intuitive dashboard displays the system health and other relevant data with D3 and Highcharts for the end user.
Technologies used: Node.js, Express.js , EJS and REST communication.

Developed a hybrid cloud infrastructure for Users to manage Mobile Sensors through a dashboard. Utilized the platform provided by NodeRed in IBM Bluemix and Rackspace’s implementation of OpenStack.

A Web Application providing complete transportation solutions in San Francisco Bay Area.
APIs used: 511 Real time transit departures API, Bay Area Bike share API, Google Maps and Instagram
Technologies Used: HTML5, CSS3, PHP, Bootstrap, JavaScript, JSON, JQuery and AJAX

A Human body browser developed using Three.js, Clara.io and other Web technologies to visually illustrate the human anatomy in three dimensions.

An E-Commerce website for buyers to advertise their products to potential sellers. The product data is sourced from Cross Domain origins using JSON data and stored in MySQL database locally, which is retrieved dynamically populated during site visits.
Technologies Used - HTML5, CSS3, Bootstrap, JavaScript, AngularJS, JSON, JQuery, AJAX, PHP, MySQL

Developed a Web based User Interface for executing Shell commands over remotely connected Unix based systems like Ubuntu, Fedora and Redhat.
Technologies Used - Java, HTML5, CSS3, JSP and UNIX Shell Commands

Developed a generic web application for conducting surveys over Geotagged images as part of The Great Mind Challenge 2011, competition. Local business owners can create surveys for their customers to answer which would help them gain valuable knowledge on their respective locality.
Technologies Used : HTML, JavaScript, CSS, JSP, Servlets, MySQL and AJAX

Developed an application for commuters to ascertain the path and the shortest route between a source and destination while being offered the choice of public transport options available as part of The Great Mind Challenge 2010 competition.
Technologies Used - Java, HTML5, AJAX, CSS3 and JSP.