Secure Cloud-Centric IR Data Management Platform

ABSTRACT: We propose to develop a secure cloud computing environment for infrared (IR) surveillance data gathered by space-based IR sensors. The IR data is typically massive and needs to be made available to warfighters via a very narrow (56k) communication channel. The current point-to-point solution of IR data compression implementation at the data downlink server reduces system"s resiliency. Therefore there is a need for distributed architecture for IR data storage, processing and management. We will design a secure cloud computing platform which will facilitate sharing of IR data and allow personalized user-developed applications by eligible users to meet their unique requirements. Our key innovation would include Google Maps-like user interface for the presentation of Geospatial IR data, and an application programming interface (API) which would provide a feedback loop to warfighters for specific spatiotemporal data requests. In order to support the narrow link, we will evaluate various image processing approaches which would minimize data transfer. Several feature extraction and motion detection techniques can be utilized for data compression which can be further tuned by user specified update threshold. We will leverage available IR datasets and the cloud computing infrastructure to evaluate the feasibility of this distributed system and user interface. BENEFIT: The Intelligence Surveillance and Reconnaissance (ISR) data captured by the space-based Infrared (IR) sensors is typically massive and needs to be delivered to Warfighters in the field via a very narrow communication channel. The cloud centric IR data platform will offer significant improvements over point to point solutions by minimizing the security attack risks, and will add great resilience and flexibility to the information system infrastructure. The cloud computing approach would allow users to request past, present and future IR datasets via web based interface with great ease and minimal management overheads. The cloud platform is envisioned as a design upgrade to existing information system infrastructure for IR data storage, processing and management. IR data cloud platform can be implemented as a stand-alone system, or alternatively it can be implemented as an extension to existing cloud infrastructures. The IR cloud platform is designed to provide access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, and services) with minimal management effort offering a low-risk, low-cost and minimally-invasive transition path to new information system paradigm. The probable commercial customers are companies that possess Big Data. The candidates are Google, Amazon, Yahoo!, Ebay, Facebook, Linkedin and Twitter. Companies such as Cloudera and MapR are also producing their own Hadoop distributions. We also see a trend in many other mid and small size companies exploiting the elasticity of the cloud. Our system will provide not only efficient cloud computing capabilities but also search, resource management and business intelligence analysis tools to commercial customers.