PDF of this tutorial
Openness
One of the most important features in programmable switching is openness. From a software standpoint, openness gives the switch developer access to switching functionality through standard programming concepts and techniques such as high-level tools with GUIs, APIs, and finite state machine definition and control. Openness also addresses adherence to industry standards—such as support for defined media resources—and network interoperability, leveraging the service provider's investment by extending the utility of the switch in a variety of environments.
From a hardware standpoint, openness lets switch developers build and configure systems through design advantages such as slot-and-device independence, component commonality, and scalability. For example, by means of a modular architecture, powerful new technologies may be enabled without the necessity of system redesign. An open switching architecture that supports industry bus standards may freely allow the optimal distribution of switching resources throughout the infrastructure, enabling developers to implement features necessary for their own unique applications. Consequently, they can bring new services to market rapidly and cost-effectively. Because of the ease and flexibility with which it can be adapted, tailored, and used to deliver multiple service applications, the open programmable switch provides a cost-effective and versatile choice for implementing network-infrastructure solutions in general.
Scalability
Scalability enables a single switching environment to fulfill a wide range of service needs by upgrading or expanding for more capacity, functionality, or performance as needed. The key benefit of a single scalable system is that it permits service providers to expand their deployments economically, efficiently, and seamlessly in step with demand. Scalable systems let providers avoid over-investment at start-up. They can commence with a small resource increment—appropriate and cost-effective for the current need—and retain capacity for later growth. When market volume and/or requirements cost-justify growth in size or scope, new capacity can be added, new resources deployed, and new services implemented as economically appropriate (a dollar-for-value investment). In Figure 4, the small switch module on the left, with under 100 ports and a footprint the size of a personal computer (PC), scales directly to the 30,000 port environment on the right utilizing the same API, switch matrix, interfaces, service resources, and architecture.
Figure 4. Illustration of Scalability
The range of scalability reflects how expandable the system is from both a physical and logical standpoint. Examples of physical limitations may be overall dimensions, number of slots, or network connections. With a modular architecture, a single distributed switch environment can support tens of thousands of ports in a nonblocking manner. This wide range of scalability extends the value of a single architecture in dynamic and growing applications and gives service providers the ability to expand resources as needed, when needed.
Scalability performance reflects the linearity expansion. If the platform's architecture is sufficiently free of performance bottlenecks, then it is possible to achieve a proportional increase in performance and capacity as components are added. In other words, an increase in capacity will result in a relatively linear increase in throughput. Ease of scalability defines how difficult it is to expand systems and whether expansion will interrupt operations. Systems can be difficult to scale or, through design competencies like hot insertion-and-removal and modularity, they can be easily scalable even while in operation. On-line expansion is particularly desirable in telecommunication voice applications, enabling services to be upgraded while the system is in operation without interrupting service.
Configurability
In a manner similar to scalability, configurability focuses on a given system's ability to be configured and reconfigured easily and flexibly for a diverse range of applications. Important aspects of configurability are as follows:
chassis design and layout—These are important for configuration flexibility. Slot, bus, and device independence within a platform chassis adds flexibility in how a hardware configuration may be structured by allowing the free intermix of boards and other functional elements with minimal restrictions on their physical location or interconnection. Compatibility with industry standards and environments can further leverage the utility of switching resources by allowing them to be deployed where and when they are needed.
modular components—Subsystems, or even entire platforms, having commonality across a variety of configurations, allow the service provider to minimize inventory, leverage equipment investment, and promote technology reuse. Architectures with sufficient granularity permit precise matching of system size to need for optimal cost-effectiveness.
performance density—If sufficiently high, this lets providers build or plan large, high-performance systems within a small footprint. Technical features such as multilayer printed circuit boards (PCBs), very large-scale integration (VLSI) semiconductor technology, midplane chassis, and superior thermal design permit configuration of more electronics per board and more boards per given space. Figure 5 is an example of a programmable switch with a midplane chassis that supports multiple switch-resource cards in a very space-efficient manner. The result is as many as 1,000 ports in a package with the footprint of a PC. This capability can greatly reduce the ancillary costs of facility development and environmental loading.
Figure 5. A Programmable Switch with a Midplane Chassis
