My introduction to Primera, which I received during the HPE Tech Event at the #NthSymp2019 was actually quite interesting. In this world of large enterprise, 5×9’s storage arrays, HPE is making a huge commitment. This is 100% uptime. Not 5×9’s, not 6×9’s, but actually 100%. Do you find this as interesting as I do? When an SLA is granted at 100%, that means that upgrades, faults, failures that are not the fault of HPE or the array (excepting things like power outages, and acts of God), I find it actually somewhat difficult to believe.
What are they doing to achieve this, though? With the integration of Infosight, the predictive analysis engine based on machine learning, this facility builds in a proactive approach toward support and the built-in ability to isolate potential problems before they exist, along with a suite of non-disruptive upgrades (most user-installable) as part and parcel to the array, much of this can be achieved. In addition, there’s a level of array based self-tuning designed around the idea that performance for every array is different in each environment, so that Primera can tune itself based on the utilization characteristics of each unique use-pattern.
Also, in keeping with the historical 3Par DNA, the Primera utilizes a large number of ASICs (Application Specific Integrated Circuits) within the architecture, including acceleration engines, and sensors for monitoring and alerting in things like performance bottlenecks, headroom, etc., the Primera is built around the concept that Tier 1 Storage has been historically way too difficult to tune and maintain, today, a Primera, feature-built with the goal of parallel performance in the disc layer, initially with SSD, and in the future, though it’s already been built to support, NVMe to come. Again, user upgradeable, and with increasingly great performance metrics.
A couple weeks ago, I was in Anaheim for the HPE Tech Summit and #TFDx held yearly during the #NthSymp2019 partner program, I had the pleasure of watching Alex Verprinsky passionately discuss the architecture, which I truly enjoyed. Here’s the video.
To me, this marks a huge step in the 3Par line. While it’s proven itself as a top-tier storage array for a decade now, the development has somewhat stagnated, up until now that is. A high-speed, enterprise class array with an SLA of zero downtime is a potential game-changer. With the densities promised today, and the increased performance and density provided by the NVMe protocol, and PCIe based NVMe based drives, the sizes and performance enhancements coming on this array are compelling. To me, this is a no-brainer in terms of considering but upgrades from an existing 3Par environment, but also in any environment that may need a greater level of performance, ease of use, and density for their tier-one data.