Several other significant intelligent storage announcements were also made, including a new Workload Planner utility for InfoSight, new Docker and Kubernetes support for HPE Cloud Volumes, the extension of Peer Persistence from HPE 3PAR to HPE Nimble – and the introduction of HPE Memory-Driven Flash, first for 3PAR, and for Nimble next year.
MADRID – “Welcome to the era of intelligent storage” was the line Milan Shetti, SVP and GM of the HPE storage business used in kicking off a press conference here at HPE Discover announcing HPE’s new announcements in the space. Intelligent storage itself is not new. Nor is intelligent storage at HPE. However, what they are announcing today expands its parameters, and will broaden its impact to more business areas than has been the case in the past.
“Intelligence is about learning from your surroundings,” Shetti said. “Thinking about it just as a destination for data is missing the boat. Storage is the custodian of data, but to unlock this potential, it has to be intelligent storage, that understands workload needs, and can adapt, self-adjust, and tune itself in real time.”
Some of the more intriguing parts of the announcement – although there are some others – are around the extension of InfoSight, the predictive analysis capability originally designed by Nimble Storage, and acquired with it when HPE bought Nimble in 2017. HPE has since expanded it into other parts of the storage portfolio. Now they are expanding it beyond storage with the addition of a cross-stack capability for providing recommendations, that extends the capability across the full stack.
“This extends HPE InfoSight to also optimize virtual machines and compute,” Shetti said. “We are now the only vendor who can provide recommendations that cover the full stack – from the compute to virtual machines and storage with everything in between.
“Customers who have had no visibility into virtual machines have identified many of them sitting idle – provisioned but not used,” Shetti explained in discussing applications of this new capability. “That’s a typical use case. They may also have overprovisioned for capacity and underprovisioned for performance, or vice versa, and this tells them how to balance better. The customers don’t need to know any of the inside stuff. The recommendation engine gives them the information. That’s the beauty of this.”
The second piece of the announcement is Workload Planner, a new AI-driven resource planner utility that helps optimize workload placement and adds the capability to run ‘what if’ scenarios on how proposed changes will impact the system environment.
“Now that you have this workload knowledge, you can use it through this new InfoSight utility to do ‘what if’ scenarios,” Shetti said. “It provides a prescriptive analysis of workload needs from Nimble storage. No one else can do a ‘what if’ scenario to see how workloads will affect the environment.”
The third way that InfoSight is extended is around performance insights capability that self-diagnoses performance bottlenecks and reduces troubleshooting time in on-prem sites that have restricted access to the cloud.
“This extends real-time analysis to sites with limited access to the cloud – like HPE 3PAR storage,” Shetti said.
Joe Ben Slivka, Manager Infrastructure and System Design at Team Motorola Solutions, then spoke to discuss how Motorola is using the full stack capabilities of HPE Insight.
“We use InfoSight to examine what we do at the storage layer, and also see if we are getting rogue VMs, to adjust on that to see if customer is getting anything on site,” he said “We can use InfoSight to pattern match what we believe we are delivering to what is actually delivered in the field.” This helps reduce outages and downtimes Motorola used to see on customer sites, and achieve a goal of less than three minutes of downtime a year.
“Anything tied to InfoSight has had no outages,” Sliva said. “In addition, this has also driven down our own costs, because it takes 4-8 hours to do a review after a customer has a problem.” He also noted that they have been using the new Workload Planner principally on the DevOps side.
Shetti also announced enhancements around HPE Cloud Volumes, a hosted storage service with hybrid and multi-cloud capability that is available for HPE Nimble Storage. This includes support for the Docker and Kubernetes container platforms.
“You now have containers-as-a-service with HPE Cloud Volumes,” he stated. “Just boot it up and you can automatically provision it, elastic scale it, and have hybrid flexibility. That is something u can do today.” Customers can register for a technical preview at https://cloudvolumes.hpe.com
HPE Cloud Volumes is only hosted in the U.S. today, but Shetti said that in 2019, global expansion will start with the U.K. and Ireland, which will make it available to customers in Europe who want local access. Shetti also announced that Cloud Volumes is now compliant with the SoC 2, GDPR and HIPAA regulatory requirements.
While much of this announcement was about the extension of the InfoSight capability from Nimble Storage, in one instance, HPE Nimble is the beneficiary of such an extension.
“One of the most requested features for Nimble has been expanding it to business continuity,” Shetti said. “Now Peer Persistence from 3PAR has been extended to HPE Nimble, which allows for multi-site synchronous replication with automatic failover. Workloads on Nimble can now failover to Cloud Volumes.” The new capability has been added at no cost to HPE Nimble Storage.
Shetti also announced HPE Memory-Driven Flash, a new class of storage built with SCM and NVMe that accelerates application performance, and marks an architectural shift to memory-driven architectures. that will drive the next generation of innovation. HPE Memory-Driven Flash lowers latency up to 2X and is up to 50 per cent faster than all-flash arrays with NVMe solid state drives.
“This is a completely different paradigm on how memory will be consumed,” he said. “3PAR will be the first platform in the industry to support storage class memory and NVMe.” That’s coming in December. It will follow for HPE Nimble at some point in 2019.
“This will come through a non-disruptive upgrade,” Shetti said.
Several expanded vendor partnerships were also announced. This includes an enhanced partnership for secondary workloads with Cohesity through an expanded resell agreement, which now qualifies HPE Apollo and HPE DL 380 servers. It also includes an expansion of the HPE GreenLake consumption-based service to include Veeam Software for backup.