What Is HCI?
HCI is software-defined infrastructure that virtualizes the elements of a traditional hardware-defined infrastructure. It evolved out of converged infrastructure (CI), which is implemented with discrete hardware that’s physically connected. HCI usually runs on commercial off-the-shelf (COTS) hardware, although it can run on proprietary hardware. The physical resources of HCI reside on a single platform for both the hardware and software layers, thus simplifying administration. This platform is usually composed of multiple servers equipped with direct-attached storage, which the HCI pools together.
The minimum components of HCI consist of software-defined storage and networking, along with a type of software known as a hypervisor that manages the virtual machines (VMs). HCI implements its software-defined elements within the context of the hypervisor. This feature combines the management of computing resources, meaning that an HCI shares its resources across all of its instances. The consolidation of an HCI’s functional elements by the hypervisor and federated management improves the efficiency of data centers and reduces their total cost of ownership (TCO).
Why dHCI vs HCI?
Disaggregation is one of the latest trends driving server architecture. Aspects of disaggregation that are beginning to appear in data centers include the decoupling of server components from a locked-down system bus, especially components. Hewlett Packard Enterprise began using the term dHCI in 2020 to indicate an architecture that delivers computing and storage in separate boxes installed in a rack-scale layout. These boxes are then connected to each other via a network fabric. The disaggregation of memory and network bandwidth is also a possible development in the future.
dHCI delivers the same experience as HCI in terms of processes such as ordering, deployment, support and management. However, it’s far more scalable, making dHCI more desirable for larger workflows. Organizations that are concerned about the performance of their HCI in larger configurations should consider dHCI because it allows them to fine-tune the ratio of computing resources to storage resources.
Benefits of dHCI over HCI
Let’s dive deep and take a look at the benefits of dHCI:
- Independent resource scaling
- Storage optimization
- Increase resiliency
- Decreased latency
Independent Resource Scaling
HCI locks users into a fixed approach to scaling that involves buying nodes and manually assigning drives to them. dHCI on the other hands allows them to scale compute and storage resources independently, which is more efficient and cost-effective. Traditional HCI is a good option for general use, but many of today’s customers want to extend HCI to more demanding applications and varied workloads. dHCI eliminates the need for organizations to choose between HCI’s simplicity and the independent scaling of three-tier architectures, which separate the user interface, logic and data processing into separate modules.
For example, users can leverage HPE Nimble Storage dHCI to take advantage of its flexible scaling for business-critical workloads. With this technology, they can add new compute hosts to independently scale within vCenter without affecting existing applications and storage. Additionally, HPE’s InfoSight system automatically detects servers and integrates them into the pool, allowing your team to add compute and storage resources independently. This capability allows dHCI to provide the best infrastructure for mission-critical data centers.
Large organizations typically need to manage petabytes of data, which is now doubling every three to five years. Optimized storage is therefore the key to successfully implementing HCI for enterprises. Their existing infrastructure requires specialized hardware such as Storage Area Network (SAN), which can be expensive to maintain. A well-designed HCI solution can eliminate the need for a dedicated storage system.
Optimizing data storage in an HCI requires the user to constantly balance performance against capacity for each VM. Achieving the ideal balance is a continual, time-consuming process. dHCI doesn’t need this trade-off because it ensures data services are always available and always optimized.
HCI users risk losing access to data and applications when a node goes down. dHCI, take Nimble dHCI for example, provides greater resiliency with its ability to tolerate three concurrent drive failures along with fast performance and high data reduction. Other solutions force the user to make trade-offs between availability, data reduction and performance by tuning resiliency, deduplication and compression. This tuning can create workload silos, resulting in the overprovisioning of resources.
Nimble Storage uses dHCI nodes with 99.9999 percent availability, data reduction that’s always on, inline compression and deduplication. These benefits require no compute overhead and increased data efficiency for all storage models. Recent improvements to Nimble Storage have further increased its resiliency with greater support for HPE ProLiant servers and integration with the latest AMD and Intel processors. Customers can also order VMware ESXi 6.7 and 7.0 pre-installed with Nimble Storage.
dHCI has lower latency than HCI, which has inconsistent communication between nodes. Organizations that want to consolidate their workloads and run performance-intensive databases require the lowest possible latency to obtain predictable application performance. In addition, they need to run inline data services without compromising the delivery of applications to their users. Administrators also want upgrades and hardware failures to be transparent to users by not impacting business applications.
Here’s a good example: Nimble Storage delivers these requirements with application performance that’s ten times faster than Dell EMC VxRail (take a look at our hpe nimble storage dhci vs dell emc vxrail blog . It also has latency that’s as low as 200 microseconds for business-critical workloads. Furthermore, Nimble Storage eliminates the resource contention issues and upgrade complexity of traditional HCI, ensuring that performance remains high. It also proactively informs customers about performance problems, along with recommended solutions.
Choosing between HCI and dHCI requires you to consider both the architecture of your data center and the scaling issues that are currently affecting your organization. HPE’s dHCI is a more scalable version of HCI that uses an integrated system of storage servers and arrays, including HPE ProLiant servers and HPE Nimble Storage. These components are integrated into a single system that uses software to automate deployment and configuration. vCenter provides the unified management that’s so important for a practical dHCI solution. Contact Comport to discuss your data needs.
Written by Tim Turner