Interconnect Driven Server
From Wikipedia, the free encyclopedia
An Interconnect Driven Server (IDS) architecture has a set of design characteristics that enable users to achieve unparalleled levels of performance, scalability, run-time flexibility and control over quality of service. System users are able to provision computing resources and communications links while being guaranteed the availability of those resources, desired bandwidth and low latency characteristics.
Contents |
[edit] The following properties define an Interconnect Driven Server architecture:
[edit] High Performance Fault Tolerant Interconnect
The cornerstone of the IDS architecture is a high performance, fault tolerant, non-blocking communications interconnect subsystem. It can be effortlessly configured and scaled to accommodate any set of applications and associated load. It provides guaranteed low latency, high bandwidth inter-processor and processor to I/O communications, and enables software driven control over all system resources including processors, I/O, communications and optional extended memory regions.
[edit] System Wide Virtualization and Quality of Service
Another important characteristic of the IDS architecture is a comprehensive virtualization framework that can set up and tear down both physical resources and virtual workspaces through explicit software commands.
Quality of service is an essential design characteristic for Interconnect Driven Server architectures. Different users, applications, workloads and their associated concurrent usage of shared communications resources can be accommodated without cross impact.
[edit] Autonomic Control
By definition, the Interconnect Driven Server architecture is self-aware of all of system components. A central registry of computing objects is managed and controlled by an independent control plane that is transparent to users. New resources that are added to the system are simply tested, put into inventory and finally provisioned through explicit software commands or policy driven business rules. Autonomic sensors proactively take suspect or failed components out of service and secure new resources from an available pool of spare components. An Interconnect Driven Server will drastically simplify system maintenance and troubleshooting.
[edit] Policy Driven Resource Management
The convergence of computing and communications creates new flexibilities and control options for system administrators. The IDS architecture allows users to set up hierarchical business rules that will proactively govern resource allocation under heavy load conditions or when component faults occur.
The IDS architecture needs to link the policy manager to the autonomic control functions. Under any fault conditions, policy driven business rules can govern the reallocation of stand-by computing or communications resources, or potentially re-allocate existing resources amongst active servers in production.
[edit] Extensibility Framework
A highly integrated communications and computing fabric opens the doors to all kinds of extensibility scenarios. The IDS architecture provides a simple and elegant means to introduce specialized resources into the underlying computing and communications fabric.
