HCI

What is HCI?

HCI stands for hyperconverged infrastructure. HCI is made up of virtual networking, administration, and storage and computing components that share resources. HCI solutions usually run on commercial,
off-the-shelf servers.

The term convergence refers to the integration of multiple technologies in a system or on a device.

A hypervisor, or virtual machine monitor, administers HCI components and it is a specialized software that manages virtual machines. Virtualization allows functions performed by a physical machine to be simulated in real time by creating a virtual version of a computer, for example, a desktop, server, or mobile device.

The main components of HCI are x86 hardware, network switches, storage capacity, and hypervisor software. Commonly, HCI clusters have a minimum of three hardware nodes to support high availability. Each node requires its own hypervisor.

What is the difference between converged infrastructure (CI)
and HCI?

HCI is an alternative to CI. One of the main characteristics of HCI is that it is software defined. In contrast, CI is hardware defined.

CI components are building blocks that can be used independently, for example a server can be removed and run separately. In HCI, all the components are completely integrated, for example the underlying server is an integral part of the hyperconverged infrastructure, and other components in the system cannot operate without it.

Unlike HCI, CI does not share storage with VMs. HCI storage architecture is composed of multiple nodes. Each node has its own storage controller service. This differs from CI storage, which is directly linked to the physical hardware.

One benefit of HCI storage architecture is faster I/O and throughput of data, mainly because there is less hardware, for example physical cables, for data to traverse. HCI uses resources more efficiently than CI because it makes data available to the VMs it manages only when they need and request it. HCI is hardware independent.

Where is HCI used? –
HCI use cases

Why is HCI used?  –
Advantages of HCI

Balanced workloads

Data centers that use HCI can easily be scaled up or down because new nodes can be quickly added or removed. This supports unplanned requirements for extra data or processing capacity when there is an unexpected surge of requests to a server. Using HCI, it becomes easier to identify real-time workload peaks and troughs because VMs only request resources from their host when needed.

Virtualized environments are isolated from each other and do not affect the workload of other guest machines. Using separate development environments for different projects promotes rapid testing and deployment of new applications.   

When HCI elements require maintenance, they can be temporarily migrated to different nodes; no separate hardware is required and there is no downtime. For users, the maintenance window is transparent, and work can continue as usual.  

HCI supports multiple business applications running concurrently. One server can virtually simulate any number of disparate systems such as messaging, networking, security, administration, finance, or new product development.

Cost efficiency

HCI benefits include lower deployment, maintenance, operational, and management costs.

HCI allows multiple virtual environments with different operating systems and functionality; in non-virtualized systems, only one operating system can be run at a time and it requires its own server. In addition, VMs in HCI systems can be built with disaster recovery options that do not need additional hardware.

The cost of HCI varies depending on the cost of the underlying hypervisor. For example, commercial solutions are more costly than open-source solutions. However, the implementation costs of open-source solutions and the potential risks associated with purchasing cheaper solutions from unknown vendors must also be considered.

Effective resource distribution and data utilization

VMs can be moved between hosts. This allows the reuse of functionality specific to a particular VM across different systems. HCI allows an organization to provide teams with specialized environments, for example for testing, development, or business-specific applications.

Virtualization enables server consolidation. Instead of purchasing multiple servers, organizations can use one server and share resources between multiple virtual environments.

HCI allows the storage of important data on the host with less important and/or real-time data stored on virtual machines.

Artificial intelligence (AI) and machine learning is used in HCI deployments to analyze the real-time use of data and to optimize storage capacity on the fly.

Simple management

HCI systems are easier to manage because all components – computing, networking, and storage – are located in the same place, namely on the host server. Centralized management increases productivity.

Unlike traditional virtualization solutions, HCI solutions enable the automatic configuration and provisioning of all system components at the same time. There is no individual proprietary software to provision separately.

HCI supports single-view analysis across a wide range of virtual configurations. This helps management to assess and to monitor systems in an organization at a high level. Modern HCI includes AI components to monitor the performance of virtual environments.

HCI workloads can easily be moved from one guest computer to another. This feature supports the rapid development and deployment of software in agile environments.

Vendor lock-in

Organizations may be locked in to a vendor if they choose a single HCI solution. An advantage of vendor lock-in is that the vendor is responsible for troubleshooting issues. Another advantage is that there may be less risk associated with a single vendor and it is easier to implement patches and upgrades.

Disadvantages of HCI

Vendor lock-in

Vendor software is often integrated with specific hardware so if an organization wants to switch to a different vendor, they may have to purchase new hardware. In addition, vendors often charge premium prices for package deals and organizations cannot upgrade individual components. 

Wasted resources

Adding additional storage nodes to an environment on an ad hoc basis can result in unnecessary memory and computing resources being added at the same time, or vice versa. When additional memory is necessary, additional storage, which may not be necessary, may also be added. This may result in inefficient scaling.

Additional costs

Virtualized workloads require large storage capacity requirements, which could result in larger volumes of redundant data. Wasted resources due to granular scaling may have additional cost implications.

Poor performance

The performance of HCI systems may be lower than the performance of CI systems. The reason is that HCI components may quickly become outdated because the testing period to verify that the complex interoperability available in HCI solutions works properly is longer.

What is the relationship between HCI and cloud computing?

The primary goal of an HCI solution is to help manage the resources that are required by different applications that are unable to simultaneously run on the same piece of hardware. The primary goal of cloud computing solutions is to limit local storage and provide remote access to data in public, private, or hybrid clouds. 

HCI can help organizations to establish a database in the cloud and it enables the creation of data from multiple applications. This data can be transferred to, and used in, independent cloud instances.

How is HCI different
to storage area
network (SAN)?

SAN provides shared storage services. SAN is characterized by block-level storage. Block-level storage simulates the behavior of traditional hard drives and can be controlled by external operating systems. Block-level storage is often used in cloud storage solutions.

SAN is different to HCI in that computing and storage functions are separate. SAN configurations need additional servers to deliver application functionality as applications do not run on a SAN storage server but require additional servers. SAN can complement HCI solutions. In this scenario, an HCI node can be preconfigured with a virtual SAN controller that handles local storage and input/output (I/0).

HCI in data centers

Data centers are increasingly moving from hardware-centric to software-centric solutions. HCI replaces traditional data center structures that have separate storage servers, networks, and arrays.

More data centers are looking to HCI, for example to cut costs. However, some players in the data center industry believe HCI solutions still lack some features. For example, HCI may not be readily equipped to deal with extreme workloads at the sub-millisecond response times that traditional data centers can.

HCI scalability, simplicity, and agility make virtualized server solutions easier to deploy and manage. In data centers, HCI can reduce overprovisioning and solve typical environment challenges like the effort and cost of maintaining physical hardware, purchasing additional servers, insuring hardware, and employing support personnel to secure and to manage hardware.

Summary

Software cannot run without hardware. HCI provides options for data centers to minimize their hardware requirements and the associated effort and labor to maintain physical data centers. HCI can also be a flexible and cost-effective foundation for the creation of virtual databases in cloud computing.

HCI technologies are able to support, not replace, cloud infrastructures where there is a massive user base, for example in social networking applications. For enterprises, HCI can be used to support on-premise private clouds. Smaller businesses can also benefit from using HCI. HCI is a way to manage multiple operating systems and disparate applications, and to employ virtualization to reduce hardware costs and increase the number of specialist environments people can work in.