The concept of computing has been around almost as long as the mathematics from which it derives its usefulness.
The world of computing as we know it today only began in the last century or so, and really only since the invention of the bipolar transistor. Fast-forward to the modern era of computing wherein we have unfathomable computing power surrounding us daily. Computing platforms and the related technology is driving advancement in every industry across the globe. From research to reality, computing platforms have been a benefactor to society time and time again. The largest benefit of modern computing platforms is that they do the heavy lifting for us – the complex calculations, the ugly mathematics, the trial and error search for a solution – without tying up a real human-being’s time and energy. Simply put, a computing platform generally is the most useful and powerful when it requires the least human interaction to complete its tasks. The ability to monitor and manage these computer platforms is as important as the jobs they perform. Additionally, as computing platforms pack more power and capability into tighter size, weight, and power constrained units, the need to easily monitor and manage these high-powered machines becomes increasingly evident.
Exploration in the Defense Industry
The Defense Industry has wholeheartedly embraced the concept of COTS (commercial off-the-shelf) tech insertion for speeding up the development and deployment of advanced capabilities to protect and defend the national interest. The wide use of COTS VPX and VME backplane technology in the Defense Industry has driven the need for a dependable platform management solution. Since many of these computing platforms are deployed in rugged and harsh operating environments, it is desirable and necessary to monitor the health of each unit in a system and manage fail-over situations should they arise. Operators are not always co-located to these systems and are not able to physically monitor and mitigate these issues. This implies that a platform management solution should be robust enough to manage fail-over situations, remotely, with limited human interaction.
High Level Solution
In keeping with the theme of COTS solutions for Defense Industry problems, the natural progression for implementing a platform management solution is to use the COTS product already in use by other industries. The Intelligent Platform Management Interface (IPMI), is the industry standard for platform management. The IPMI architecture “defines standardized, abstracted interfaces to the platform management subsystems” . The IPMI specification describes a message based request/response protocol in which commands are grouped by functional sets using their Network Function Code or NetFn for short.
|04||Sensor and Event Request|
|05||Sensor and Event Response|
This protocol can be initiated over a wide variety of interfaces such as a serial port, a Keyboard Controller Style (KCS) interface, or an Intelligent Platform Management Bus (IPMB). The serial port, KCS interface and other interfaces comprise a class of interfaces known as System Interfaces. System Interfaces are used by the unit itself to communicate with its onboard IPMI controller. The IPMB belongs to a class of interfaces which are used to communicate with a unit remotely; i.e., not from within an application level software on the unit itself.
These commands over an IPMB are used to monitor and manage the individual units in a larger system. In IPMI terms, each unit would be called a Field Replaceable Unit or FRU. An example of a FRU might be one Single Board Computer (SBC) in a VPX chassis. Network Function Codes exist for managing sensor data, reading and writing FRU data, configuring and acknowledging events and alerts, powering on or off a FRU, and more.
The IPMI spec also allows for extensions to its standard yet robust set of commands. One extension is the VITA Standards Organization’s VITA46.11 Specification . The VITA46.11 spec implements a set of states and commands on top of the IPMI specification which allow for more granular control of the FRU by defining FRU states, as well as a set of standard logical sensors which summarize the FRU health.
A device implementing these standards on a FRU is called an Intelligent Platform Management Controller or IPMC. Baseboard Management Controller (BMC) is another term used to describe an IPMC. The VITA46.11 spec divides implementations into Tier I implementations and Tier II categories. A Tier I IPMC implements a subset of the standard, whereas a Tier II IPMC implements the entirety of the standard.
Pairing the IPMI standard commands with VITA46.11 additions, an IPMC can be used to monitor and manage a FRU in a wide variety of situations. The real value in having an IPMC onboard a FRU, is that when multiple FRUs exists in a chassis together, an IPMI compliant Chassis Manager has access to an incredible wealth of information about what is occurring in the system. For example, in a system with multiple FRUs which contain IPMC v2.0 compliant IPMCs, the Chassis Manager can instruct each FRU to monitor certain voltages or temperatures and send an alert known as a Platform Event Message if an out-of-range event occurs on the FRU. Not only can the FRU be instructed to monitor itself, but the Chassis Manager can request any and all sensor data at any time from any FRU. This is a valuable asset in the view of system management.
The need to monitor and manage computing platforms grows with the increasing complexity and power of those platforms. In the Defense Industry, COTS products are a solution to quick and standards compliant tech insertion with respect to computing platforms, and platform management is no different. Using standards compliant IPMCs will reduce or remove the need for custom platform management solutions and ease system management greatly.
 IPMI v2.0
 VITA 46.11