AdvancedTCA targets net management

If the AdvancedTCA platform provides the basis for developing various telecommunications systems for next-generation network infrastructure equipment, what additional capabilities will be built on top of that platform to achieve the level of management automation traditionally expected in the central office?

The answer to that question lies in how effective AdvancedTCA will be in allowing the construction of a modular communications platform. The platform will only consist of standard building blocks, including hardware, an operating system, standards-based interfaces, and middleware and application services that support customer applications on a high-availability system.

In the hardware layer, the AdvancedTCA platform provides for broad redundancy of components and interconnects (i.e., redundant compute and I/O, redundant or virtual networking, RAID and redundant power supply). Further, the platform will remove single points of failure in the system, thereby enabling significantly high availability.

To complement a highly available hardware platform, the industry needs a hardened carrier-grade operating system (CGOS) that is validated and designed to run on AdvancedTCA hardware. A CGOS would be designed to run on AdvancedTCA. Broad-based hardware support enables the same applications to run on different hardware systems without requiring significant changes.

A CGOS software platform is based on a kernel with the hardened features and improvements to broadly cover the following areas: reliability, availability and serviceability improvement and capabilities; kernel hardening and improvements; and hooks to support system monitoring and management middleware.

The high-availability management middleware is a key element. It is essentially an implementation of high-availability features in software designed as an extension of existing management middleware to provide a number of essential features. Those features include availability management and real-time data collection; system modeling and dynamic system configuration; fault-management capabilities; and automated replication and fail-over features, including support for both stateless and stateful applications.

To make this high-availability-management middleware effective, the CGOS offers lower-level features that include redundant CPU fail-over; Ethernet fail-over; event-logging and resource-monitoring capabilities; clustering framework; intelligent peripheral management interface (IPMI) system management; and an application programming interface for hardware failure prediction.

Telecommunication applications can be developed directly on top of the CGOS interface (using various interfaces) or on the middleware interface-and it is possible to use both approaches together. Examples of customer applications include control applications targeted for call processing, media gateways, soft switches and applications for operaional support system-related services. Those applications allow service providers to provide services to their end customers.

The AdvancedTCA platform addresses backplane fabric implementation in stages that relate to interoperability with current equipment, evolving technologies and industry-driven standardization to support interoperability. In concept, the AdvancedTCA platform will allow the interconnection of various new elements of the network by incorporating fabrics standardized by the industry. Those fabrics will evolve within a standards-based platform that is capable of supporting the equipment requirements expected over the next 10 years.

In the PICMG 3.0 spec, PICMG defines the base interface as Ethernet 10/100/1000 Base-T, with addi-tional specifications for the fabric interfaces in PICMG 3.1 for Ethernet, PICMG 3.2 for InfiniBand, PICMG 3.3 for StarFabric and PICMG 3.4 for PCI Express. Other fabric interface implementations might be specified in the future.

One of the most important considerations for the development of a modular communications platform has been system management. To establish a hardware platform that can meet the requirements of the management system, it is necessary to understand the management capability for tomorrow's equipment.

One requirement of management will be monitoring platform resources and components: shelf modules (such as cooling fans, fan trays and air filters); power resources; boards installed in the shelf; and environmental conditions (such as plant power, voltage and temperature). The AdvancedTCA shelf defined by the PICMG 3.0 specification provides for such features. Typically, the shelf manager provides instrumentation for health metrics and alerts when the reporting indicates such out-of-range conditions as low voltage or high temperature.

Management interoperability requires standardization of the platform and application management interfaces. PICMG and the Service Availability Forum are examples of industry organizations driving standards.

Other requirements in management automation might be derived from a system manager that also accesses an AdvancedTCA board's management middleware to obtain information like board performance. The middleware may use that functionality for the purpose of board fail-over in case of a component failure within a shelf.

To support management automation overall, however, the middleware could report to a system manager the fail-over conditions, resources affected by the fail-over and the resulting successful continuation of services.

Health and performance metrics from a shelf manager and system manager can determine the board resources-such as bringing additional shelves of equipment online-needed to support a system that is providing services.

Advanced deployment systems could script installation of multiple applications in successive stages, which could include operating system, drivers, management middleware, services such as Signaling System 7, a billing application or network services. The implementation of such flexibilities would be determined by the complexity of the software deployed, the necessity for staged deployment and the time available to deliver software to the board.