10 Gigabit Ethernet Options for the Data Center

Design Editorial

Tuesday, 23 September 2008 15:27

Written by Ms. Kohl

As a result, most new data centers are now being located based upon the cost of energy, or said another way, the cost of cooling.  However, many existing data centers must try to optimize and live within the boundaries of their current facilities.

Regardless of the scenario, and although cost is certainly an issue, it is not enough to think only of the corporate “pocket book”.  All power sources have some type of detrimental effect on the environment, and therefore, it is incumbent upon everyone to minimize the negative impact on the environment while maintaining alignment with corporate objectives.  There are a number of ways these issues can be addressed including software as well as improved hardware implementations.  These include virtualization, upgrading to higher networking speeds in the data center, and optimizing for the cost/performance associated with these higher speeds relative to the usage requirements of the given data center.

The Savings and Costs of Virtualization

Virtualization is one of the key tools being used to help maximize the utility of a data center.  Server virtualization increases the utilization of servers by permitting one physical server to operate as many virtual servers.  The reduction in the number of physical servers saves on real estate costs, lowers electrical bills, and saves on cooling expenses.  There is minimal power differential when operating a CPU at 10 percent utilization versus operating one at 70 percent utilization.  However, reducing the number of physical servers to power, and cool, can result in significant cost savings.

The Development of 10GbE for Volume Applications

Although 10GbE is a much discussed technology, historically its deployment has been largely restricted to wide area networks (WANs), metropolitan area networks (MANs), high-performance computing (HPC) and selected enterprise deployments.  The reasons for its limited deployment have been primarily related to cost, power, performance, and “user friendliness” issues. 

Almost all 10GbE deployments to date have used either traditional optics modules combined with fiber cabling or short reach copper cabling using a CX4 cable.  These technologies lack the ability to drop down to lower speeds; and therefore, the addition of 10GbE into a data center required a forklift equipment upgrade.  The 10GbE network also lacked the ability to run at lower speeds and use less power.  

Another issue associated with 10GbE has been the limited performance of the 10GbE networks.  This problem is not related as much to the 10GbE technology as it has been to the microprocessors, host bus, chipsets and memory interfaces that were used in server platforms. Often times, microprocessors and chipsets consume and dissipate significant amounts of power yet lack the efficiency needed to utilize network bandwidth.  New, state-of-the-art microprocessors and chipsets are more efficient in handling the network bandwidth, but still consume most of the power budget in a server platform.

Lastly, the historical cost of these solutions and cabling ease of use issues have not made 10GbE an attractive solution unless the increased data traffic demands cannot be met otherwise.  For example, CX4 is constrained by its lack of ability to field terminate cables, offers a poor bend radius, and is limited to a reach of 15 meters which has made it difficult to use in many applications.  However, through the ratification of new standards and developments in technology, the ability to use 10GbE in server platforms made great strides during the course of 2006 and into 2007.10GbE is now proving to be a viable, cost-effective and energy-efficient technology.

Deploying 10GbE Using Small Form Factor Pluggable (SFP+) Optics

In 2005, work began by the Small Form Factor (SFF) Committee to develop a specification that would reduce the size and cost of optical modules as well as offer power savings.  The project, now know as SFF-8431, is developing the electrical specification for interfacing to cages and modules defined in SFF-8432 and SFF-8083.  The SFF-8431 specification defines the electrical interface for 8.5 and 10 Gbps SFP+ modules and hosts.  The module is a hot pluggable small footprint serial-to-serial data-agnostic optical transceiver intended to support data communications applications.  This specification provides a common reference for system manufacturers, system integrators and suppliers. 

As of January 2008, work on this specification is still on-going with a targeted ratification date of mid-2008.  Pre-standard products are now available in the market and the risk associated with adopting SFP+ is now minimal.  Many vendors are already embracing SFP+ as an attractive alternative to previous 10GbE optical module offerings due to the increased port density.  In addition, as an industry average, power is reduced by approximately 20 percent and cost is reduced by about 10 percent.

Check out the November print issue of the Data Center Journal for the complete article from Blaine Kohl of the Ethernet Alliance, including a look at deploying 10GbE over Copper Cabling, Real Estate Requirements, Server and Switch Implementations, Reducing Routing Complexity and Energy-Efficient Ethernet in the Data Center.

About the Author and the Ethernet Alliance
Ms. Kohl is the VP of Marketing for the Ethernet Alliance, a non-profit industry consortium and sits on the advisory board of 4Blox, a software iSCSI company.

The Ethernet Alliance is dedicated supporting the education and expansion of Ethernet-based technologies.  For more information on the Ethernet Alliance, visit www.ethernetalliance.org.

TrackBack URI for this entry

Comments (0)

Subscribe to this comment's feed