September 1, 2014
I’ve recently been thinking a lot about application-specific workloads and architectures (Optimize Scalalable Workload-Specific Infrastructure for Customer Experiences), and it got me to thinking about the extremes of the server spectrum – the very small and the very large as they apply to x86 servers. The range, and the variation in intended workloads is pretty spectacular as we diverge from the mean, which for the enterprise means a 2-socket Xeon server, usually in 1U or 2U form factors.
At the bottom, we find really tiny embedded servers, some with very non-traditional packaging. My favorite is probably the technology from Arnouse digital technology, a small boutique that produces computers primarily for military and industrial ruggedized environments.
Slightly bigger than a credit card, their BioDigital server is a rugged embedded server with up to 8 GB of RAM and 128 GB SSD and a very low power footprint. Based on an Atom-class CPU, thus is clearly not the choice for most workloads, but it is an exemplar of what happens when the workload is in a hostile environment and the computer maybe needs to be part of a man-carried or vehicle-mounted portable tactical or field system. While its creators are testing the waters for acceptance as a compute cluster with up to 4000 of them mounted in a standard rack, it’s likely that these will remain a niche product for applications requiring the intersection of small size, extreme ruggedness and complete x86 compatibility, which includes a wide range of applications from military to portable desktop modules.
At the other extreme, we have systems for extreme high-end database and analytics, which have been steadily encroaching on the territory that even as recently as half a decade ago had been the protected turf of high-end RISC-UNIX machines. We have seen a stream of announcements of 4-socket and even a few 8-socket systems based on the latest Xeon E7 V2 CPUs from Intel. While these are impressive systems, with about double the throughput per socket as the previous generation and memory footprints of up to 1.5TB per socket for extreme in-memory analytics and database operations, scalability seemed to have hit a wall, at least as far as vendor’s willingness to innovate in a limited market …
… at least until I talked to Bull SA recently about their “bullion” (sic – no caps) system. Bull, a highly qualified French company that is essentially invisible in the North American market, has been making high-performance computers for several decades, working with various US vendors such as Honeywell and IBM, and learning the craft of making very scalable mainframes for critical applications. The bullion servers seem to validate their status as master craftsmen at the high-end. The bullion servers scale to a full-rack configuration of 16-sockets, currently the largest Xeon production servers on the planet, and at 240 cores and up to 24 TB of memory, they represent the largest monolithic x86 platform for in-memory processing available, and one that comes close to the flagship configurations being highlighted by IBM and Oracle, probably at a fraction of the price.
With the announcement cycle for Intel’s new Haswell midrange CPUs and accompanying mainstream hardware partner products looming, I fully expect that the ingenuity of the extended x86 ecosystem will be on display over the next few weeks, but I didn’t want to lose sight of the extremes in what will certainly be a tsunami of coverage on the next core enterprise product cycle. Of these two priducts that I have selected to highlight here, probably the bullion server is the one that might actually be of interest to most Forrester clients and readers of this post – Bull may actually be in a position to expand their very modest N. American footprint with a server that is twice as scalable as any other server in the market and four times that of any US-based vendor.