You’re reading an outdated article. Please go to the recent issues to find up-to-date content.
At its June 2001 meeting, the TelecOMMunications Industry Association (TIA) Engineering Committee TIA TR-42 approved formation of a new working group, TR-42.1.1, for network distribution nodes. As is often the case in standards organizations, when you cannot reach agreement on a commonly used term, you create a new one—network distribution nodes. This a term was created because this standard is envisioned to encompass many different facilities of IT. Some common industry terms the proposed standard will address include: • Conventional data centers—main frame networks • Customer-owned intranet and Internet data centers • Storage area networks (SANs)—both owned and leased by the enterprise customer • Service distribution nodes or storage service providers—third-party facilities for rent • Internet server farms The formulation of this new committee and the eventual development of a standard recognizes that cabling is often overlooked, but making the right decisions and investment initially will save time, money, and frustration in the near term and will minimize disruptive upgrades in the future. The Internet data center (IDC) or storage area network (SAN) cabling needs differ from those of a traditional cabling network. The agreed-upon issues for the group include: • Current commercial building standards do not accurately reflect the needs of the emerging Internet data centers (IDCs) • Deployment time and impact of the existing infrastructure and customers must be factored into the design • The driving force behind the network design will be density and optimization of allocated space • IDCs have unique requireents for path redundancy and customer connectivity Here are some personal observations. The ability to reliably handle huge data streams of information at the fastest rate is critical. No one can afford for the system to pause for hours, minutes, or even seconds while transferring the information. It has to be only a “click” away. Fast deployment is of the essence, especially for third-party operations. Servers on a pallet or a nonoperational facility are not producing revenue or receiving orders. Ideally, the cabling components are modular, simple to use, and pre-tested. Revenue is too valuable—particularly for start-up companies—to wait weeks, days, or even hours for equipment to be installed and cabled up. High density is critical because the number of connections is tremendous and amount of equipment is enormous, while the cost of environmentally controlled, fire-protected, redundant power, high-security facilities is high. The cabling must allow for flexibility and reduce the cost and time necessary to make moves, adds, and changes to the network while also offering redundancy. Lastly, it must allow for a low initial capital investment, especially for the cabling, providing a system that can be installed in phases on an as-needed, as-generating-revenue basis. The scope of the working group will include topologies (cabling design) and performance for copper and fiber cabling, and other aspects of the IT infrastructure that will enable these facilities to rapidly deploy existing and new technologies. It is most likely that the cabling recommendations will be based on supporting the highest data rate applications such as Fast Ethernet, 1 Gbps Ethernet, 10 Gbps Ethernet, as well as 100/200/400 Mbps Fibre Channel. The committee recognizes the need for a rapid development of the standard and has established an internal goal to have the first revision completed within 24 months. A very rough outline of topics to be addressed has been developed and includes topology, cabling distances, recognized cables, choosing types of cabling, cabling installation practices, and grounding considerations. Given the need to support the highest data rates, this standard will certainly be pro-fiber, probably with a bias toward 50-micron fiber for short distances within a single facility and single-mode fiber connecting facilities. Given the density requirements, not only will small-form-factor (SFF) fiber connectors be allowed, but they will probably be required. Additionally, the use of higher-density fiber connectors, such as the MPO, which is a 12-fiber connector the size of an RJ-45, will be discussed. The need for density combined with the MPO connector will lead to significant discussion and recognition of higher-density ribbon-constructed fiber cables. Because of the need for rapid deployment and ease of installation, the use of pre-connectorized cable assemblies versus field-installed cable will be recognized. Lastly, given the need for flexibility and the fact that IDCs are open floor spaces with either raised floor or overhead pathways, the use of a structured form of zone cabling is almost certain to be adopted. TIA TR-41.1.1 has many issues to reach consensus on; however, expertise and experience does exist and will be drawn upon based on past lessons learned from conventional data centers and the earliest implementations of IDCs. BEAM is director of systems marketing at AMP NETCONNECT Systems. He can be reached at (336) 727-5784 or [email protected]