In this article

Over the last several years, the server and storage industry has given way to several hyper-converged infrastructure (HCI) products. Most of these products are very similar in nature, by bringing software-defined storage (SDS) products to the modern data center.

To this end, in 2019 Hewlett Packard Enterprise (HPE) launched the disaggregated HyperConverged Infrastructure (dHCI) platform. Internally named ProStack, dHCI brought to market the flexibility of a converged stack with the manageability of HCI.

This stack of servers, storage and switches consists of HPE Proliant Servers, Nimble Storage and a variety of approved options for storage switches including Aruba, Cisco, Mellanox and others.

HPE Nimble Storage DHCI 2.0

The HPE dHCI system allows the customer to be up and running in as little as 15 minutes with the help of the dHCI specific software that comes pre-loaded on the Nimble array. Additionally, the customer is able to scale compute and storage independently of each other, creating the feel of scalability afforded traditional converged architecture.

A major advantage to the deployment of a dHCI stack is the ability to integrate with HPE InfoSight. InfoSight is an artificial intelligence (AI) driven tool that can help predict potential issues within the environment, open support cases and drive remediation. The dHCI solution shows up in InfoSight as a single solution, tying together servers and storage for faster support resolution and more direct performance monitoring.

When deploying the HPE dHCI solution, a minimum of two HPE Proliant servers running VMware ESXi (see the documentation for specific models) is required and can scale to 32 VMware ESXi server nodes (larger scaling to be introduced in later releases). Once you have deployed the two initial servers, the customer can add more in any increment up to the platform maximum. In addition, unlike some hyper-converged solutions, the customer can add non-ESXi servers to the Nimble array within the stack to run physical workloads.

The Nimble SAN that is part of the dHCI solution is like any other Nimble SAN with the exception of one thing — the management software that enables dHCI. Full Nimble OS functionality exists within this solution and replication tasks can easily be set up between a dHCI-enabled Nimble and a non-dHCI Nimble array.

Initial releases of the HPE dHCI stack are relegated to iSCSI. However, they include both Hybrid Flash and All-Flash options. Adding storage to the solution is a snap and no different than adding a shelf to a traditional Nimble SAN. The customer would purchase a shelf unit to attach via redundant SAS ports on the array, and the storage space would increase with zero downtime.

Increasing the number of HPE Proliant nodes in the solution has never been easier. Ordering the Proliant server as a dHCI enabled resource allows it to come preloaded with VMWare ESXi from the factory. This facilitates an easy "plug and play" during install. The server is discovered via service location protocol (SLP) from the vCenter plugin in vSphere. Once discovered, the end user can select the server, populate required TCP/IP information, enter passwords and the node is ready to go.

Bringing this level of simplicity and scalability to the data center is the primary target of the HPE dHCI solution. Combining the manageability of an HCI stack with the expandability of a converged infrastructure is the best of both worlds for the customer data center. This solution transforms the division of traditional workloads on physical servers and virtual workloads on dedicated virtual solutions from the isolation of running separate systems.

Get started

For further information on Nimble, including labs, automation and general information, take a look at the resources in our ATC.

If you're interested in learning more about converged infrastructure, check out our HCI workshop or learn more from the WWT experts on how the ATC can help.