Open standards converge to close on new markets
Many open standards that originally stood alone are now being successfully combined with other open standards to create very powerful platforms that, while becoming very sophisticated and powerful, are taking less time and money to develop and deploy. This cross connection of open standards happens because the people who are involved with creating them figure out ways to put standards together to create something bigger, better, faster, more flexible, and cheaper. The theme of this issue is how standards are being combined and expanded, and how the results are creating new market opportunities not originally envisaged by their creators.
A dozen or so years ago, telecom equipment suppliers realized that they needed to stop building computers and devote their resources to their core intellectual property, which is communications technology. To this end, three standards that are increasingly being used together are xTCA (which defines a very scalable hardware layer and the platform management that controls it), Linux, and the High Availability (HA) management software defined by the Service Availability Forum (SAF) that sits between the hardware and Operating System (OS) and the application layer at the top. The combination of this powerful trio of standards is changing the way designers think about high-reliability systems, which are critical to more and more markets beyond communications.
A match made for multiple markets
My friend Dr. Asif Naseem provides a high-level overview of this triad and explains how the three cornerstones of a modern platform work together and are increasingly being adopted by Communications Service Providers (CSPs) and Network Equipment Providers (NEPs). COTS hardware and sophisticated HA middleware not only make systems highly reliable, but improve maintenance and upgrade strategies over the lifetime of equipment in large systems. Mark Overgaard of Pigeon Point Systems dives more deeply into the topic of SAF-compliant platform management and how practical implementations work. Mark chairs a PICMG committee that develops standards to this end, and last year completed two Hardware Platform Management (HPM) specs ratified by PICMG, HPM.2 and HPM.3.
Interestingly enough, HA systems like those enabled by xTCA combined with carrier-grade Linux and SAF-compliant HA middleware are coming to the attention of military systems designers, who are architecting their platforms more and more like telecom systems as communications between sensors, vehicles, aircraft, and command centers becomes vital to battlefield situational awareness. In the military space, high reliability used to mean a heavy box milled from a block of aluminum that contained electronics that were heated, frozen, and shaken to indicate reliability. While these measures improved reliability somewhat (and cost a lot), earlier architectures still had what are commonly known as "single points of failure" such that if a critical component died, the system ceased operating. This meant that determining the failure required the equipment be removed, shipped to a repair depot somewhere, dismantled, and eventually returned for service. The concept of HA, which monitors all system elements and can switch in redundant backups as necessary, means that critical systems continue to work in the event of a hardware failure or software fault. Military designers are embracing this approach more and more, and it is interesting to note that this is an example of a concept used for many years in the telecom space migrating to the military world. Steve Looby from SANBlaze is interviewed in this issue, and he explains how the HA concept is being readily accepted by developers of military electronics.
HA is also moving into "Big Science" applications, specifically High-Energy Physics (HEP). Markus Joos from the European Organization for Nuclear Research, better known as CERN, describes in his article the very demanding nature of particle accelerator detector instrumentation, with data channels numbering in the hundreds of millions and time resolutions measured in picoseconds, and how CERN is moving towards xTCA as an overall architecture for the future. It's another world. CERN has led the effort to extend several PICMG standards (specifically PICMG 3.8 and mTCA.4) for physics instrumentation, and PICMG will shortly be releasing a Design Guide for timing, clocks, and gates in these applications.
Closing on shares
In closing, a recent market study done by IMS Research shows that xTCA and CompactPCI together comprise about 50 percent of the embedded board market now and will do so through at least 2016. This research report was presented at the recent Embedded Tech Trends Conference and can be found at opsy.st/M2AMarketResearch.