Following up on our recent HCI blog (link), we wanted to update you on the latest developments. Over the past decade, we have come to realise the benefits of HCI (Hyperconverged Infrastructure)- the simpler, software-defined ability to create VMs en masse, manage them, and move them around, as needed, regardless of the hardware. However, there are some drawbacks. Basic HCI, while still perfect for VDI, and general-purpose virtualisation, can’t always handle more complex workloads that are needed to scale. For instance, even if you could put line-of-business apps like ERP or transactional database environments into a VM, you need reliable infrastructure that minimises failure domains so that one node outage does not create downstream issues for critical applications at scale.
Now, what if you could have an HCI environment where you could scale to maximum performance, and reduce the risk of losing apps and data, and remove wasted resources tied up in multiple backups? Well, there is such a solution – our technology partners over at HPE have developed the HPE Nimble Storage dHCI, short for disaggregated HCI, also called HCI 2.0. HPE Nimble Storage dHCI is an intelligent platform with the flexibility of converged and the simplicity of hyper-converged infrastructure (HCI).
The platform disaggregates compute and storage and integrates hyper-converged control, providing simple infrastructure management on a flexible architecture. Built with the world’s best-selling server, HPE ProLiant, and the self-managing flash storage of HPE Nimble Storage, this platform provides the flexibility to scale compute and storage independently for unpredictable growth, and the data resiliency and performance needed for business-critical applications. Powered by HPE InfoSight, the HPE Nimble Storage dHCI platform is intelligently simple, absolutely resilient, and efficiently scalable- this last part is unique in that it performs that “needle in a haystack” job, to find, report and sometimes proactively fix things that have, or will soon, go awry.
A LOOK UNDER THE HOOD
Resilient design center for 99.9999 percent availability (with Nimble Storage) with no single-point-of-failure and advanced data integrity to tolerate three simultaneous drive failures. Low-latency performance- as low as 200 microseconds data response with automatic QoS to ensure fast performance for every application. Industry-leading data efficiency with advanced data reduction and a modern OS providing up to 21x data reduction.
With HPE Nimble Storage dHCI, you can run faster with rack-to-apps in less than 15 minutes, avoid disruptions with the help of predictive analytics that delivers 99.9999 percent data availability with HPE Nimble Storage, and can optimize with higher productivity and maximum resource efficiency:
Intelligently simple with native, full-stack intelligence from storage to virtual machines (VMs) and policy-based automation for virtualized environments:
- Features include a fast, self-service experience of hyper-converged control with simple setup and auto-discovery for the entire cluster via VMware vCenter.
- All data services are integrated with VMware vSphere and VMware virtual volumes for a native VM experience.
- Also included are what-if simulations that eliminate guesswork when consolidating new applications, as well as app-aware recommendations for self-optimizing performance and resources.
Resilient and ready for demanding apps on a platform designed for 99.9999 percent availability with HPE Nimble Storage and sub-ms of low-latency at consistent, high performance:
- Features include data-centric visibility extending across the infrastructure and every VM. Performance bottlenecks are easily diagnosed with root causes from storage to VMs. Advanced data services, such as QoS and synchronous replication, help ensure applications stay on and are always fast.
Efficiently scalable to eliminate overprovisioning and lower TCO compared to competitive platforms. Features include the ability to independently scale compute and storage non-disruptively with industry-leading data efficiency guaranteed.
A STEP BY STEP GUIDE TO dHCI
Deploy the HPE Nimble dHCI solution by going through simple steps with the help of dHCI setup software. The steps include:
- Physical infrastructure layout
- Complete the configuration worksheet
- Network configuration
- Initialize and configure the HPE Nimble Storage
- Deploy a new vCenter or use the existing vCenter
- Add HPE ProLiant to the HPE Nimble Storage dHCI stack
- Create VMFS and/or VVOL datastores
vCenter Plugin, After the deployment completes, you can perform various tasks from the vCenter plugin for HPE Nimble Storage dHCI. Adding a new server to dHCI can be accomplished with these steps:
- Rack the server and perform the cabling
- Configure the switch and assign the correct VLAN
- Assign an IP address to the ESXi management interface
- Add the server in the vSphere cluster using the vCenter plugin
In the background, dHCI software automates tasks such as the configuration of vSwitch, iSCSI software initiator, and VMDK binding during deployment- saving additional cycles for administrators. Utilise the dHCI plugin to create and grow VMFS and VVOL-based datastores that remain mapped to volumes on HPE Nimble AF/HF arrays configured as part of the dHCI deployment. Additional operations, such as clones and snapshots on the datastores, can also be performed directly from the HPE Nimble dHCI plugin on to vCenter.
A CENTERPRISE SUMMARY
Part of the beauty of this solution is- simplicity. It’s simple to deploy – taking just 15 minutes from rack to application. It’s simple to manage – using the VM-centric vCenter tools you’ve grown to know and love. It’s simple to scale – adding storage and compute with no disruption and no impact to the apps, and you can even add existing HPE ProLiant G9/Gen 10 servers you already have. It’s simple to support – HPE Infosight monitors and alerts complex issues, so you don’t get surprised. And you can have this solution and pay for only what you use, with HPE GreenLake for Nimble Storage dHCI.
See some blogs on this from our colleagues over at HP: