More Internet bandwidth is continually needed due to the general expansion of Internet usage and to the adoption of bandwidth-hungry applications, including IPTV, video, peer-to-peer sharing, and wireless backhaul. Computer and network architecture and technology have kept pace with such advancements as multi-core processing, virtualization, networked storage, and I/O convergence. These improvements have similarly led to increased demand for data center bandwidth.
Historically, increased bandwidth requirements have been addressed with solutions that begin by combining multiple lower bandwidth links, and then transition to higher bandwidth links. We've seen this multiple times, in the transition from 10 Mbps to 100 Mbps to 1 Gbps to 10 Gbps.
Today, both within data centers and across major Internet links, multiple 10 Gigabit Ethernet (10 GbE) links are aggregated using one of several link aggregation protocols. Aggregation, however, comes with a raft of problems. Use of multiple 10 GbE links requires a large-scale replication of resources, including expensive computer interfaces and switch ports. Aggregation of multiple flows can be a complex operation, calling for packing of flows that use less than 10 Gbps of traffic and splitting of flows that use more than 10 Gbps. The bursty nature of most Internet protocols, coupled with stringent SLAs, often results in bandwidth wastage.
It's essential that single, logical, high-speed links be available in the data center and throughout the Internet. The step to 100 GbE is seen as a key ingredient to the next major expansion of the Internet - "the core empowers the edge."
If these high-speed links are ever to become ubiquitous, standards and conventions must be developed. The IEEE has taken the lead in 40 GbE and 100 GbE standards. Starting with the inception of the IEEE HSSG (high speed study group) in mid-2006, the Task Force (IEEE 802.3ba) was formalized at the end of 2007 and is scheduled to issue approved standards by 2010. At the same time, multiple network component vendors are participating in multi-source agreements (MSAs) for internal and interface components not dictated by the standards.
The members of the IEEE committee strongly believe that 100 GbE is ready now - that no new breakthrough is needed to proceed. In fact, due to the immediate need for high-speed links, NEMS and carriers have implemented the 802.3ba pre-standards. There's likewise an immediate need for test equipment to be used by NEMs, carriers, and enterprises to validate full line-rate operation, verify interoperability, and validate protocol operation.
To meet the needs of 100 GbE users, Ixia has concentrated its efforts toward providing IEEE 802.3ba test equipment that tracks the evolving standards for 100 GbE.
Ixia's 100 Gigabit Ethernet Proof of Concept Demonstration offers:
- An IEEE 802.3ba Task Force-based implementation
- The first 100 GbE line-rate traffic generation and analysis solution
- The first PCS-layer implementation using multilane distribution (MLD)
When available, Ixia's 100 GbE interface card will merge seamlessly with all other Ixia interfaces and test applications. As a compatible member of Ixia's test platform, it will allow full layer 2-7 testing with all types of interfaces, at all speeds.
For more information, including white papers on 100 GbE and PCS/MLD, please go to http://www.ixiacom.com/100GbE. For a private demonstration and briefing at NXTComm'08 - e-mail Thananya Baldwin, Strategic Programs Director, at tbaldwin@ixiacom.com to make an appointment to see this leading-edge technology.
Tags:
No comments:
Post a Comment