Dinesh Singh

Hitachi Hyperconverged Infrastructure – Top Ten Q and A

Blog Post created by Dinesh Singh Employee on Jul 17, 2018

Hitachi Vantara  Unified Compute Platform HC (UCP HC) hyperconverged systems, powered by industry-leading VMware vSAN software, has been around for over one and a half years. Here is a sampling of the most common questions .

 

1. Why Is UCP HC ready for prime-time?

UCP HC has been successfully deployed in real-world, revenue-generating tier 1 applications  Customer experience illustrates high performance, reliability and availability for demanding production workloads. Customers initially began with smaller, single purpose installation have since expanded capacity and/or added additional, workloads. We have several public customer references from various geographies deploying a wide array or workloads.  Read how a South-African retailers Dis-Chem Pharmacy here is leveraging UCP HC for retail

AAA.png

2. Can UCP HC scale to support large IT environments?

UCP HC can scale up to 64 nodes in one cluster, resulting in  petabytes of storage capacity. It offers linear scale as you add more nodes. Customers can design multiple clusters that can all be managed with single instance of vCenter. Adding capacity is a seamless exercise as the software automatically detects the new node and rebalances the cluster by distributing the storage objects in compliance with your availability policies. Best practice for designing UCP HC cluster can be found on Hitachi Knowledge Base.

Picture2.png

3. Can Hitachi’s external array Virtual Storage Platform (VSP) systems be attached to UCP HC ? Why do you want to do this?

UCP HC can be connected to external-controller based storage VSP. This design is supported by Hitachi and VMware engineering teams. VMDKs can be distributed across UCP HC datastore or VSP’s VVol datastore(s) and/or VMFS datastore(s). This enables two main use cases.  First, seamless migration of workloads from legacy environment (including baremetal) to UCP HC, where the data was already on the VSP array.  Second, this allows additional storage expansion without being forced to immediately upgrade to another set of hyperconverge nodes in your cluster, if the environment is capacity-bound, as opposed to compute or memory-bound.

Screen Shot 2018-07-17 at 2.55.38 PM.png

4. How do I modernize data protection with UCP HC?

UCP HC enables you to select a combination of three schemes of data protection to meet your availability needs

Business Continuity –

Within cluster - VM administrators define the protection level for VMs through policy, while creating the VM. Availability level is expressed in terms of FTT (failure to tolerate), which is defaulted to 1, meaning two copies of data across nodes. Natively built-in protection ensures availability during any failure of disk, switch, node or rack. Protection against network partitions is done through a quorum voting algorithm.  If the network partitions for some reason, >50% of the votes for a given storage object is needed to allow reads and writes to proceed.  This protects against split-brain syndrome.

   a. Replication – UCP HC nodes can be designed in metro storage configuration (~50 mile) to ensure high availability despite the site failure during area wide natural disaster. For cross-geo (beyond 50 miles) protection, vSphere replication can be leveraged to create asynchronous copy of data.

   b. File data (especially for ROBO and VDI deployments) can be stored offsite at Hitachi Content Platform (HCP) at central location. This integrated solution from Hitachi delivers low-cost, scalable infrastructure with multiple level of data protection, while ensuring compliance with organization’s IT security policies.

Operational recovery –

Customers can store data copies outside the HC cluster using Hitachi Data Instance Director (HDID) for recovery purposes. HDID accelerates application consistent copies and ensures faster recovery to rebuild the data. Reference architecture to protect your hyperconverged data with HDID can be found here. It provides VADP backup for UCP HC.

Long-term retention -

Customers in industry segments which require business data be stored for long-term for complying with government or industry regulations can benefit from Hitachi’s customized solution, bringing together UCP HC, HDI and HCP. Have written more about this solution here.

 

5. What is Hitachi’s strategy with NVMe technology on hyperconverged platform?

Lately, NVMe has been a buzzword in enterprise storage that promises to dramatically improve I/O performance for higher consolidation ratio and improve latency for critical apps. Recognizing the promise of NVMe storage to address the needs of extremely-low-latency workloads, Hitachi introduces NVMe drives in UCP HC systems in 2017. NVMe technology is evolving rapidly and Hitachi will continue to bring innovations in this transformative system to the market.

 

6. What are new enhancements with UCP HC to help customers address current and future HCI needs

Since inception, UCP HC has been on rapid release cycle to add new features and functionality that make it a reliable platform for a wide range of virtualized and container workloads. Recent release in June 2018 included the following key enhancements:

a. Faster application performance with Intel Optane NVMe – After introducing NVMe SSD for cache system in fall of 2017, we have introduced Intel Optane NVMe, which is equipped with 3D XPoint memory, and designed for high performance and faster responsiveness. It’s generally available now. Use cases such as real-time data analytics, high-performance database and high-performance computing will significantly benefit from NVMe

b. High-end graphics applications (Graphics Processor Unit) - NVIDIA GPUs provide the computational  engine for the new era of visual workspace, enabling amazing user experiences by accelerating deep    learning applications at scale. Newly announced UCP HC system with Tesla accelerators deliver the    horsepower needed to run bigger simulations faster than ever before. For enterprises deploying VDI, Tesla    accelerators are perfect for accelerating desktops to any user, anywhere. Tesla GPU-enabled UCP HC    V225G will be the ideal platform for collaborative design developments and for performing CAE simulations for crash analysis, heat and pressure simulations in automotive development scenarios.

      c. Accelerated access to data with high network speed – Data access is now faster than ever with

      25Gbps. Customers have multiple options of configuring the high network speeds with 10 and 25Gbps.

 

7. How will UCP HC integrate with my existing IT infrastructure?

UCP HC leverages vSAN software-defined storage, which is embedded in the vSphere hypervisor for higher performance and minimal impact of CPU and memory. vSphere is predominant virtualization technology and has a broad ecosystem of management software for virtualized storage and networking built around it. Many existing common Hitachi and VMware customers chose to go with UCP HC because it blended with their existing environment without introducing new management layer.

 

8. How do I transform IT to hybrid cloud operating model using UCP HC?

Hyperconverged architecture is offers accelerated path to enable hybrid cloud with just-in time deployment flexibility and elasticity to provision or decommission capacities in an uninterrupted manner.. For self-serviced, metered, template-based resource consumption, VMware vRealize Automation cloud management software can be leveraged on UCP HC systems. It provides native integration with vRealize Orchestrator to automate routine IT operations with workflows. UCP HC provides hyperconverged building block for integrated turnkey hybrid cloud solution Unified Compute Platform RS

Screen Shot 2018-07-17 at 10.36.46 AM.png

9. Are UCP HC and VDI perfect match to deliver native desktop experience and lower operational cost ?

For IT shops getting started with virtual desktop infrastructure (VDI) transformation or exploring the replacement of legacy three-tier infrastructure, UCP HC helps modernize the VDI infrastructure by helping you avoid the guessing game of right sizing compute and storage capacities. UCP HC is simpler, easier to manage and less expensive to buy and operate. Icing on the cake is the optimized pricing for Horizon VDI licenses, if it’s being deployed on UCP HC, substantially bringing down the cost of VDI project. If you’re considering VMware Horizon WorkSpace One, UCP HC is simply a no-brainer. You can read more on best practice document for deploying Horizon VDI on-premises here and with the flexibility and pay-per-service consumption model here.

 

For Customer’s who are running large business-critical VDI environment, manageability and cost saving is key. They often struggled with sizing external storage as the number of users grew. However, with UCP HC, it’s as simple as creating policies for performance, protection and security and underlying software manages the rest in the background removing all guess work from VDI administrator.

Picture6.png

10. How do I manage hardware element on UCP HC systems?

Hitachi Unified Compute Platform Advisor (UCP Advisor) is the life-cycle management software for UCP HC series. It is also life-cycle management tool for other members of UCP family, rack-scale systems UCP RS and converged infrastructure UCP CI, simplifying element management for customers deploying multiple UCP systems.It allows firmware management for compute and network components. UCP Advisor provides integration with vRealize suite of cloud automation software. It allows vLAN provisioning to networking ports. It provides visibility into hardware and software, besides power cycling and LED management for nodes. For customers having experience with UCP Director, UCP Advisor modernizes the infrastructure management with similar comprehensive functionality as UCP Director, with added flexibility of modular pricing.

 

If you have any questions around UCP HC, please post them in comment section below.

Outcomes