With the rise of in-memory computing, it turns out that eight-socket x86 servers are now one of the fastest growing classes of infrastructure technologies in the data center. Big Data analytics applications that need to be processed in real time in particular lend themselves well to eight-socket servers that allow applications to more economically scale up.
With that in mind, Hewlett-Packard Enterprise today unveiled a new eight-socket x86 server based on Intel Xeon E7-8800 v3 series processors. Dubbed the HPE Integrity MC990 X Server, it can be configured with up to 6TB of memory. The HPE Integrity MC990 X Server makes all that memory addressable using SGI interconnect technology that HPE has licensed to provide a coherent shared memory architecture that makes it possible for as many as 144 processors in the system to access memory directly.https://o1.qnsr.com/log/p.gif?;n=203;c=204663295;s=11915;x=7936;f=201904081034270;u=j;z=TIMESTAMP;a=20410779;e=iIn general, Jeff Kyle, director of product management for enterprise servers at HPE, says HPE is applying x86 server economics to a class of servers that used to be dominated by RISC processors. The goal is to not only reduce the total cost of computing, but also make it easier to meet almost any level of application service required in an age when massive amounts of data are becoming routine.
At the same time, Kyle also notes that thanks to lower memory costs, the overall physical footprint of data center environments is shrinking as memory replaces magnetic disks as the mechanism for providing primary storage. That not only means denser server environments; the amount of energy being consumed by those data centers will be substantially less.
The degree to which those capabilities influence where Big Data applications get hosted in the future remains to be seen. But given the fact that accessing those applications in a public cloud introduced additional network latency, chances are the majority of these applications will continue to be deployed on premise for some time to come.