Industry certifications – such as NEBS (Network Equipment-Building System) – are important to ensure the quality and reliability of telecom equipment. NEBS is the most frequently used guideline of its type for telecommunications equipment in the United States.
The standard ensures, too, that communications equipment can continue to work in any type of situation. Telcordia (News - Alert) manages NEBS specs. AT&T, Verizon, BellSouth, and Qwest (News - Alert) formed a Telecommunications Carrier Group, which refined NEBS standards. For example:
• NEBS Level 1 has low threshold for equipment hazards and network degradation. It is primarily used for getting prototypes into lab trials.
• NEBS Level 2 is used in connection with equipment operability in controlled environments, such as datacenters. It is used rarely.
• NEBS Level 3 has specs for fire suppression, thermal margin testing, airflow patterns, acoustic limits, earthquakes, failover and partial operational requirements , failure severity levels, RF emissions and tolerances, and testing/certification requirements.
NEBS criteria apply to many NEI-certified reference platforms. So too do other industry certifications, such as FCC (News - Alert), UL and CE.
There are other items which apply to NEI-certified reference platforms, according to a recent NEI white paper on reliability issues. These include having enough power for normal operation of components; redundant, hot-swap power with failure alarming; tested Redundant Array of Independent Disks (RAID) subassemblies that include high performance SCA SCSI, and SAS (News - Alert) hard drives, custom designed hot-swap hard drive carriers, RAID controllers, and cabling; stable operating system platforms and kernel versions; components based on industry standard chipsets, leading CPUs, memory, backplanes and peripherals; industrial rack-mount enclosures; and NEI (News - Alert) reference platform certification.
There are also some basic tasks to ensure reliability. These include:
• Do not install multiple applications on a server to then use the server for multiple functions. When it comes to servers, a single system generally has a single purpose.
• Do not install untested, uncertified or unsupported apps on a server that has been pre-configured.
• Do not install third-party utilities, programs or device drivers that directly affect hardware, especially in operating systems that use a hardware abstraction layer. They may cause system failure.
• Review app log files and event logs for errors or application status. Write up reports on available disk space, memory and CPU utilization.
Proper maintenance is also important to ensure server reliability. Some suggested types of maintenance include responding to hardware alarms and make repairs quickly; responding to alarms from software agents; using UPS, battery backup and/or industrial surge protection; do not operate a server in a questionable environment, for example, a room with high humidity; and change or clean enclosure filters.
Ed Silverstein is a TMCnet contributor. To read more of his articles, please visit his columnist page.
Edited by Jamie Epstein