Today, Hitachi announced the availability of a new midrange all NVMe storage platform (VSP E990) which provides some eye-catching outcomes for VMware-Hitachi infrastructure customers as foundation for their next SDDC deployments. Existing customers can expect to see 2.5x IOPS, 3-6x lower latency and significant parallel performance advancements in our adaptive data reduction capabilities (ADR).
With respect to the VMware ecosystem, I recently posted a blog discussing Hitachi's support for the new VMware vSphere 7.0 release and some of the key use cases that we were focused on. The VSP E990 is certified and supported with vSphere 7.0 on day 0 and of course 6.x releases. The Hitachi-VMware vRealize/vCenter integrated management portfolio for VSP Storage will also extend to E990 in a phased certified rollout in time for production deployments. But, I do want to expand on some of those use cases I referenced in the other blog and share new ones now that we have super charged VSP E990 in the portfolio.Virtualize and Modernize Legacy with vVols and Storage Class Memory (SCM)One of my personal favorites with the SVOS powered VSP E990 is the array virtualization capabilities combined with VMware vSphere vVols and a Storage Class Memory (SCM) ready E990.The VSP E990 can non-disruptively front-end (virtualize) non-NVMe or legacy arrays (including 3rd party arrays) to allow them life extension or non-disruptive migration enablement while gaining the NVMe performance, resiliency and capacity efficiency benefits of the platform. To add to this, as insightful customers move with pace to using next-generation storage with VMware vSphere Virtual Volumes (vVols), the legacy arrays inherit the advantage of our unique storage policy based management (SPBM) and VASA/vVols implementation that we have delivered.As an example, apps and data tend to age in importance over time but never go away. How can you effectively manage this at scale at a VM level and give App owners that control for their services. Simple, by providing a portal to simply assign a vCenter VM storage policy with capabilities of Tier 3 IOPS+Latency for these VM services, our VASA provider will detect that policy change and automatically tier those warm/cold VM/VMDKs to the lowest legacy tier in this configuration freeing up that tier 1 for the next revenue generating app. The beauty is the vSphere admins are managing just ONE vVol datastore (storage container) across this diverse storage footprint as we have abstracted two/multiple storages into that one vVol datastore using our Hitachi data tiering (HDT) technology. They no longer have to know or manage 100’s of datastore belonging to different storage targets. This tremendously simplifies operationalizing this efficiency outcome for ops and apps teams managing their infra allocation.I baited the SCM ready nature of E990. Think of a Tier 0 in this configuration being storage class memory and still presenting one vVol datastore. Stay tuned for more on that brilliant innovation soonVMware Cloud Foundation (VCF) and external storage
This 56 core beast with 1TB of cache, 100% data availability guarantee and NVMe backend encryption translates to an outcome to consume more VM and container workloads in smaller footprint, deliver better effective capacity with impressive security, performance and availability SLA's. The target price point and utility pricing model options are one of the other eye catching items for me giving the capabilities offered leaving those ROI calculators dizzy!
We continue to see strong interest in VMware Cloud Foundation (VCF) with external SAN storage. The use cases vary by customer ranging from flexibility to scale storage footprint, mission critical applications with stringent RPO/RTO requirements to simply matching business outcomes to suitable tier of storage. With VCF 4.0 expanding their external storage offerings for principal/supplemental storage with support for vVols / VMFS to augment native vSAN datastores, I foresee the E990 NVMe platform being a popular choice augmenting vSAN. The E990 can deliver the same performance of 200+ SAS SSDs with just 40 NVMs drives, driving that smaller footprint.
Virtual Desktop Infrastructure (VDI)
Hopefully most are aware that Hitachi provides a VCF powered appliance offering using our Hitachi Unified Compute Platform (UCP RS) including lifecycle management of compute/network/storage with UCP Advisor which will be ready to be augmented with E990.
We have been successful in supporting businesses with their VDI infrastructure using our popular vSAN based UCP Hyper-Converged offering (UCP HC) offering in combination with VMware Horizon View. But what sometimes gets lost in the noise is that we have a multitude of customers where they have significant scale and desires to consume SAN storage as a known operational model to deliver VDI infrastructure with Citrix XenDesktop and VMware Horizon View.
Kubernetes Containers and Persistent Storage :-We recently rolled out an architecture guide for using persistent cloud native storage (CNS) for containers (vSAN, vVols, VMFS and BM) and this extends to VSP E990. With the growth in containers, we foresee the need for persistent storage for what I believe will be predominantly Cassandra and more so my favorite PostgreSQL databases as targets for those container services. These services will access their ‘persistent data’ through shared but probably sharded databases instances. With that foresight, we have done significant OLTP performance benchmarks with VSP E990 and believe we are well positioned to meet these VM/container workload profiles.
One large customer is running over 10,000 Horizon VDI desktops with current VSP storage across 151 compute nodes. With a typical knowledge/power worker requiring anywhere between 13-37 IOPS per desktop of predominantly write I/O, they won't get to see the published hero number of 5.79M RRH IOPS when upgrading to VSP E990 for this workload, but will see significant increase in their desktop count as they move to Phase 2 for additional users. They can also leverage the spare capacity for traditional server workloads.
Additionally, customers can now take advantage of the additional work we have done to provide storage policy management for persistent storage for Kubernetes clusters (including Tanzu Kubernetes guest clusters) running on top of vSphere to deliver the requested storage class capabilities at vmdk level.There is a lot more to speak about and I will provide additional updates around Ops Center automation services (hint, Ansible tech preview ready storage modules), advances in three-data-center business continuity solution with VMware Site Recovery Manager and new data protection options with Ops Center Protector in subsequent blogs. Don't forget to check out other blogs/FAQs around this innovative announcement. If you have specific questions, feel free to reach me @pmorrissey3000 on twitter or leave comment here.