IDT PCI Express(r)在有线通讯系统中的作用

IDT PCI Express(r)在有线通讯系统中的作用:USING THE IDT PCI EXPRESS STANDARD
SYSTEM-INTERCONNECT SOLUTION TO
DELIVER PREDICTABLE, SECURE, HIGH-RAS
COMMUNICATIONS IN BLADED SYSTEMS
The bladed system communications challenge
Bladed system architectures enable equipment providers to produce high-density, modular
equipment that can be easily scaled to match user needs. The inherent flexibility and scalability
of a bladed architecture supports the expansion or reconfiguration of existing server, storage
and telecommunications infrastructure, and supports the rapidly growing and shifting
services demand.
This high level of scalability is achieved by disaggregating computing and scarce system I/O
resources. However, it poses a significant challenge: the system needs an equally scalable
interconnect architecture that enables any and all root complexes to communicate both with
each other and with any and all endpoints.
The IDT™ system-interconnect solution, the industry’s first PCI Express® (PCIe®) system-interconnect
solution, fulfills this demanding requirement. Developed in close collaboration with leading customer
design teams, the solution is optimized specifically for demanding bladed system designs.
The four minimum operational requirements of a system-interconnect architecture are that it must:
• Be easily expandable to support extensive system scalability
• Deliver predictable, wire-speed performance under all load conditions, regardless of system
form factor and capacity
• Ensure high reliability, availability, serviceability (RAS) and security
• Consume low power
However, system complexity mandates a level of functionality that simplistic peer-to-peer switching
cannot deliver. The complete solution must:
• Support unconstrained inter-processor communications
• Support maximally configurable system architecture
• Enable the flexible virtualization and sharing of peripherals and scarce system I/O
• Dynamically map I/O and peripherals to computer resources
• Facilitate inter-domain communications in systems that deploy heterogeneous processors
• Reuse existing hardware, software and firmware wherever possible