As a result of the real-time nature of voice communications, the concept of reliability in switching is all-important. Reliability is a key factor in switching cost-effectiveness. The more uptime that a given investment in hardware and software produces, the greater potential for revenue and profit as a result of that investment. Reliability has two primary metrics:

availability—This is a measure of the ability of the switch or switching environment to remain continuously available for service—in operation—without interruption. Typical availability statistics are calibrated in mean time between failure (MTBF), usually hours, or as in uptime as a percentage of total time. The standard for select, highly reliable programmable switches is in the vicinity of 99.995 percent or greater.

maintainability—This measures the ease and speed with which operation can be restored, should an interruption occur. It is often expressed by mean time to repair (MTTR), usually in minutes. Sophistication can cut MTTR to microseconds via an automatic and high-speed cascade of events: error detection, isolation of failed components, and switchover to backup components. Maintainability is also extended by features that enable carriers to service or even expand equipment without interruption of operations. For example, the additional ability to hot-plug (physically remove, replace, or add components while the switch remains in operation) can also enhance maintainability, as it eliminates the need to power-down the system.

Beyond pure numbers is the potential ability of the switch to tolerate a fault in a manner that is transparent to operation. It is generally enhanced by building in redundancy throughout the switch architecture. In redundancy, an extra component is incorporated into the system. It serves as a powered backup to the covered component. In case of a fault, it can rapidly and automatically be brought into service. Redundancy can be applied to virtually any level in the switching environment, including individual buses and chip components, switching matrices, circuit boards, and etc. In distributed architectures, entire switch chassis, resource node, and fiber-ring redundancy can be provided. Full 1+1 redundancy implements a one-for-one backup of each component and is commonly employed for critical functions served by a single component such as a switch matrix, signaling controller, or fiber ring. In applications such as network interfacing, where a collection of numerous identical components are deployed in a load-sharing arrangement, N+1 redundancy can be a viable and cost-effective alternative by providing one back-up component per collection.

In summary, an open programmable switch gives service providers a resource that is as reliable, configurable, and scalable—from a hardware standpoint—as it is open from a software standpoint.

Price Performance

Programmable switches can offer significant price performance advantages over earlier-generation switching alternatives. Through distributed and scalable architectures, system resources such as processing power, bus bandwidth, and memory can be applied where they are needed. Sophisticated switching techniques distributed in service resources can further conserve ports for network operations. Because switching resources can be sized, configured, and programmed to suit the application, high performance can be achieved cost effectively.

Standards Compliance

By its very nature, programmable switching strongly implies adherence to standards. Standards in programmable switching apply not only to the openness of the switch environment itself, but also reflect compatibility to the standards required by the network infrastructure that the switching function is serving.

Network standards

Because the switch is a functional element in the infrastructure, it must comply with a variety of functional telecommunications standards such as those that govern ISDN, SS7, T1, and E1. Compliance with country-specific standards, protocols, and signaling variants are required for certification within a given locality. Examples include the American National Standards Institute (ANSI) and International Telecommunications Union (ITU) standards, the European Union's connection endpoint (CE) Mark, and a multitude of individual international standards and deltas such as the British Approvals Board for Telecommunications (BABT) in the United Kingdom and Japan Approvals Institute for Telecommunications Equipment (JATE) in Japan.

Regulatory requirements

In addition to telecommunications standards, the switch must generally comply with other technical specifications such as electrical safety, heat output, radio-frequency emissions, and electrostatic discharge. These are published by government or industry regulatory bodies such as new equipment building system (NEBS) (Bellcore) and Underwriter's Laboratories (UL).

Device standards

In order to integrate host processing, switching, and media resources such as voice-processing devices, standards have been established to support their related bus architectures. Current de facto standard bus architectures include signal computing system architecture (SCSA), multivendor integration protocol (MVIP), and PCM expansion bus (PEB); broader, industry-standard buses are likely to emerge in the future. Such standards enable a variety of board-level devices to be compatible across different platforms. Recently, the industry has taken a step further and begun to embrace interoperability between various architectures and APIs. A leading example of this effort is the Enterprise Computer Telephony Forum (ECTF), an industry organization whose family of interoperability agreements (IAs) are intended to accelerate deployment of computer telephone integration (CTI) technology. For example, the H.100 specification has been proposed by the ECTF as an industry-standard bus for media-service resources.