For highly scalable database, line-of-business, and custom applications
Windows Server 2008 for Itanium-Based Systems offers a foundation optimized for the most compute-intensive and critical business analytics and enterprise applications.
Windows Server 2008 for Itanium-Based Systems provides a fault-resistant platform suitable for high-end RISC/UNIX system replacement, floating-point intensive applications, and applications that share data extensively among processes.
The hot-pluggable component support and simplified clustering in Windows Server 2008, when coupled with the reliability, availability, and scalability (RAS) features of the Itanium platform, enable customers to confidently move proprietary RISC/UNIX and mainframe implementations to a standards-based architecture with improved agility and total cost of ownership (TCO).
Intel Itanium Processors
The Intel Itanium processor features native 64-bit processing coupled with the Intel Explicitly Parallel Instruction Computing (EPIC) architecture.
Itanium does not extend the 32-bit x86 architecture to handle 64-bit computing—instead, it features a completely new instruction set optimized for 64-bit processing. The result is high performance for compute-intensive server workloads that require a highly scalable platform and the capability to directly access more RAM and a larger virtual address space.
The cornerstone of the Intel Itanium architecture is EPIC technology. EPIC increases processor performance through instruction-level parallelism, which strives to maximize opportunities for parallel execution, making more efficient use of on-chip resources. The Itanium compiler spends more time optimizing the instruction flow for the processor, helping the final executable run more efficiently.
Mainframe-Class Reliability and Availability
Succeeding in the global marketplace requires remote employees, international customers, and partners to have around-the-clock access to information. A disruption in services can result in diminished productivity and lost opportunities.
Features of the operating system, server hardware, and the Intel Itanium processor architecture provide increased reliability, availability, and scalability (RAS) for workloads running on Windows Server 2008 for Itanium-Based Systems.
Dynamic Hardware Partitioning
Windows Server 2008 for Itanium-Based Systems supports your server hardware, so that you can take advantage of dynamic hardware partitioning and make changes to your server partitions on the fly:
- Hot Add Memory
- Hot Replace Memory
- Hot Add Processors
- Hot Replace Processors
This helps increase the reliability, availability, and serviceability of servers; you can replace a processor that shows signs of failing or add spare processors to a partition as demand increases, all without restarting the server.
Failover Clustering
Windows Server 2008 for Itanium-Based Systems supports clusters of up to eight nodes. Failover clustering provides broader high-availability capabilities for business-critical workloads. The marginal cost of adding failover clustering is low, since you can have many cluster nodes running your application with only a single node standing by for failover.
Itanium RAS Features
- Windows Server 2008 for Itanium-Based Systems makes full use of RAS features at the heart of the Intel Itanium processor architecture:
- Memory Error Correction Code (ECC) enables detection and correction of single-bit data errors in processor caches and data bus.
- Enhanced Memory ECC retries double-bit errors.
- Memory Sparing predicts a failing dual in-line memory module (DIMM) and copies the data to a spare memory DIMM.
- Memory Mirroring writes data to two locations in system memory so that if a dynamic random access memory (DRAM) device fails, the mirrored data enables the server operation to continue to run.
- Memory Cyclic Redundancy Checking (CRC) automatically retries address and command transmissions if a transient error occurs.
- Lockstep Support, in conjunction with supporting hardware, enables redundant processors to simultaneously execute the same instructions with the same data, allowing for recovery in the event of data corruption.
- Advanced Machine Check Architecture (MCA) with Extensive Error Correcting Code (ECC) provides more advanced error-handling capabilities (such as correctable double-bit errors in the table look-aside buffer [TLB]), with well-defined interfaces at the hardware, firmware, and operating system levels.
- Symmetric Access to All Processors enables a server to restart and operate if the primary processor fails.
Massive Scalability
As organizations rapidly grow and change, IT infrastructures must remain flexible and scalable. They must help support, manage, and secure expanded functionality, an influx of additional users and locations, and increasingly robust applications.
Windows Server 2008 for Itanium-Based Systems scales in the following ways:
Processors and Memory
Windows Server 2008 for Itanium-Based Systems supports up to 64 cores (for example, 32 dual-core processors) on a single server, and up to 2 terabytes of RAM.
Next-Generation TCP/IP Stack and Networking Components
Windows Server 2008 for Itanium-Based Systems increases scalability and performance changes and enhancements to the core networking components, and lets you migrate to IPv6 as the rest of your networks do:
- Receive-Side Scaling (RSS) allows a network adapter on a high-volume server to receive more traffic.
- Receive Window Auto Tuning enables better throughput between TCP peers and lets you increase the utilization of network bandwidth during data transfer.
- Policy-Based Quality of Service (QoS) for enterprise networks lets you create policies to ensure greater QoS for certain applications or services that require prioritization of network bandwidth between client and server. In this way, you can help keep network resources from becoming bottlenecks when scaling mission-critical applications.
Mission-Critical Performance
The high performance of Windows Server 2008 for Itanium-Based Systems comes from the features of Intel Itanium processors:
EPIC Technology
The Itanium EPIC architecture provides superior Instruction-Level Parallelism (ILP), resulting in higher performance and faster time to insight. (Up to six instructions can execute per clock cycle.)
Large Memory Cache on the Processor
The 24 megabytes of on-die L3 cache provides low-latency data to both cores on dual-core Itanium processors. This increases data throughput and transaction processing.
Large Number of Processor Registers
Itanium processors have 128 registers.
Summary of Features
Windows Server 2008 for Itanium-Based Systems offers the following high-end features and capabilities:
- Support for 2 terabytes of RAM
- Scalable to up to 64 Itanium processors (or 64 cores)
- Hot Add/Replace Memory and Hot Add/Replace Processor support
- Eight-node failover cluster support
- Fault-tolerant memory synchronization
- Licensing rights to run an unlimited number of virtual instances of Windows Server 2008 for Itanium-Based Systems with third-party virtualization technology
|
