Sélectionner une langue
Red Hat Enterprise MRG was launched at the Red Hat Summit in June 2008, providing customers with a next-generation IT infrastructure that offers increased performance, reliability, interoperability and faster computing through the combination of Messaging, Realtime and Grid functionality.
Today, version 1.1 of Enterprise MRG was released. With 1.1 comes enhanced features for each of the components of Enterprise MRG – Messaging, Realtime and Grid. MRG’s Grid technology was available only as a Technology Preview in MRG 1.0, but today is fully supported and production-ready for those with a Red Hat Technical Account Manager (TAM).
New highlights and features in Enterprise MRG include:
- Messaging: native infiniband and RDMA driver for dramatically better latency, clustering, enhanced security, queue semantics like Last Value Queue and Ring Queue, Native .NET client and improved management tools.
- Realtime: improved performance, especially on boxes with higher CPU-counts, and improved performance tools. For example, Tuna now has the ability to write tunings to an init script once you’ve found an optimal tuning for your system.
- Grid: enables enterprises to achieve higher peak capacity and IT utilization through the creation of high-performance grids by leveraging their existing infrastructure. It provides the most advanced and scalable platform for both High-Throughput Computing (HTC) and High-Performance Computing (HPC), and also offers the ability to schedule to all available computing resources including:
- local grids,
- remote grids,
- virtual machines,
- idle desktop workstations and
- dynamically provisioned cloud infrastructure.
With Red Hat Enterprise MRG customers are offered:
- Messaging, up to 100-fold faster than competing solutions.
- Faster response times.
- The ability to schedule to local grids, remote grids, virtual machines and rented cloud infrastructure; and cycle-scavange from idle PCs to use all available computing resources.
- Reduced complexity and breakthrough value through an integrated platform.
- Deterministic performance. MRG Realtime provides high levels of predictability for consistent low-latency response times to meet the needs of time-sensitive workloads.
- Deterministic low-latency messaging. The combination of realtime and messaging functionality enables messaging with highly deterministic response times and reliably low latency.
- Scalable computing power and resource utilization. Schedule workload to local grids and remote grids, cycle-steal from desktop workstations, and dynamically provision and schedule work to cloud-based infrastructure like Amazon EC2. MRG also supports virtualization to harness available resources for various jobs.
- Messaging-based reliable, low-latency job scheduling. By leveraging MRG’s Messaging and Realtime components, MRG Grid can reliably schedule and execute jobs in less than a second.
- Scalable and efficient utility computing. Enterprises can move toward a utility model of computing where applications are scheduled as jobs across a heterogeneous set of computing resources. MRG’s Grid scheduler can then automatically optimize utilization, priorities, load, and other key factors.
- Large-scale virtualization. MRG provides sophisticated capabilities for managing large-scale virtualization deployments across heterogeneous servers with changing workloads.
- Unified management. Red Hat provides a unified management platform with a Web-based console, command-line tools, and API’s across the entire MRG platform - greatly simplifying the complexity of managing distributed computing.