Platform MapReduce is the best-of-breed, next generation distributed runtime engine for enterprise –class Hadoop MapReduce applications. Platform MapReduce is compatible with many popular MapReduce distributions. Unlike distributed workload engines found in open source and other commercial MapReduce distributions, Platform MapReduce is designed to provide enterprise-class MapReduce runtime capabilities by delivering high resource utilization and predictability, high availability, an open architecture supporting multiple applications and file systems, better manageability and enterprise –class security. Platform MapReduce is built on Platform Computing’s years of expertise in distributed workload scheduling and resource management capabilities, both are proven technologies that are powering many Fortune 500 companies for their mission critical, most demanding workloads. As a best-of-breed solution, Platform MapReduce delivers unprecedented distributed workload runtime services for your MapReduce applications.

Product Architecture

Platform Computing’s MapReduce solution includes (see architecture diagram below):

• Application adapter technology for executing Hadoop MapReduce jobs without the requirement to change code or recompile. Support includes MapReduce Java, Pig, and Hive application code. In addition, other Hadoop projects such as Oozie are also supported.
• Multiple Application Programming Interfaces (APIs) for other commercial application execution. Examples include R, C/C++, C#/.NET, Java, Python, R, native binaries & others.
• Support for mixed application workloads (MapReduce and Non-MapReduce APIs) executing on the same set of shared resources within the same cluster.
• Platform MapReduce Workload Engine that automates, distributes, and manages MapReduce workloads according to users’ Service Level Agreements (SLAs).
• Platform Resource Orchestrator that allocates and manages distributed pools of resources including clusters, servers, CPUs and memories. It allows multiple applications to share a common set of resources.
• Advanced file system and data access framework – Provides connectivity to different types of file systems and database architectures, eliminating the need to migrate existing data while optimizing resource utilization.
• Rich set of management, troubleshooting, and reporting tools – Single GUI interface for managing and troubleshooting multiple MapReduce applications across a shared set of resources. Full version compatibility with support for rolling upgrades.