SA Forum HPI and AIS Specifications
By Asif Naseem, Service Availability Forum/GoAhead Software

Service Availability Forum (SA Forum) is a vendor consortium dedicated to specifications for service availability middleware. Its two key specifications so far are:

• The Hardware Platform Interface (HPI) allows a user application to identify available hardware resources.
• The Application Interface Specification (AIS) establishes ways for a user application to utilize the middleware.

The SA Forum specifications provide three key benefits:

• The ability to discover, monitor, and manage a platform’s hardware resources in a generalized way.
• Portability of applications developed using AIS across middleware versions.
• Standardized access to management capabilities of HPI and AIS services.

Hardware Platform Interface Specification (HPI)

• Discovery: HPI allows user applications to discover and enumerate a system’s hardware resources and their management capabilities. Such resources could, for example, include compute cards, networking boards, and power supplies. The HPI service collects and maintains information such as the manufacturer ID, product name, and serial number.
• Monitoring: The HPI service includes detailed mechanisms to monitor the hardware resources’ health and performance. The notification and logging capabilities provide a way to monitor, communicate, and log events. Events could include a temperature exceeding a threshold or a server failure.
• Management: Manageable hardware resources called HPI entities generally have instruments associated with them, such as sensors, controls, watchdog timers, and annunciators. Instruments provide information, read/write access to control devices, mechanisms for customizing the communication of alarms, and indicators of events.

Application Interface Specification (AIS)

The AIS consists of an availability management framework and eight services that middleware can use to develop highly available systems and services. The services are:

• Availability Management Framework (AMF) : Coordinates redundant resources to eliminate single points of failure in a cluster.
• Information Model Management (IMM) : Provides a way to represent, access, and manage system components in an information system model
• Cluster Membership (CMS) : Maintains up-to-date cluster membership information as nodes enter or leave the system.
• Checkpointing (CKPT) : Provides the capability to record and retain dynamic state information. A redundant resource can use that information to seamlessly resume the service provided by a failed one.
• Event (EVT) : provides a mechanism for applications to subscribe to receive events as and when they occur.
• Messaging (MSG) : provides an efficient mechanism for communicating information such as application states, event and error notifications, and fault management data.
• Logging (LOG) : Enables network/system administrators (or automated facilities) to review logged information to troubleshoot issues such as misconfigurations, network disconnects, and unavailable resources.
• Notification (NTF) : Provides a means by which applications can send notifications related to system events to interested entities.
• Lock (LCK) : Provides a distributed lock management capability that resolves conflicts over shared resources.

Conclusion

Many vendors have already implemented both HPI and AIS. The SA Forum plans to release updated versions with new features and enhancements in January 2007.

Asif Nassem is a Member of the Board of Directors of the Service Availability Forum and Chief Operating Officer, GoAhead Software.