We are willing to bet that most people do not know or have little idea what the word “hyperconvergence” means. It did not exist two decades ago, but today it is an important part of building IT infrastructure in both large and relatively small companies. In this article, we tell you what hyperconvergence is, what are the advantages of such systems for the consumer, how you can properly use its appliances, and whether you can mix them.
Hyperconvergent is a kind of infrastructure in which all the traditional server hardware elements are combined into a single space and orchestrated using software. It turns a complex of servers into one computing system, which is pretty easy to manage through a common console. To scale the system, it is enough just to add the required number of new nodes.
How hyperconverged infrastructure works
The software that is used to build a hyper converged infrastructure allows you to abandon the traditional storage system. Instead, local hard drives of the servers themselves are used. Software tools work according to the basic principles of modern storage systems – fault tolerance, integration into arrays of independent disks, and others. In addition, they provide an even distribution of the available internal drives. For example, if three servers are combined into a hyperconvergent infrastructure, then there will also be three mirrored copies of the data.
Hyperconvergence significantly increases the efficiency of server virtualization: it becomes more flexible in terms of scaling, easy to manage and fault tolerant. This is the optimal solution for medium and large businesses, as well as small companies that want to save space in racks and are ready to pay for the required number of software licenses.
Hyperconvergent infrastructure software
An important component of the hyperconverged infrastructure is software for system deployment and management. The software is crucial for implementing hyperconverged infrastructure based on the high-performance architecture. It allows you to scale to meet the future needs of the company, reduce the total cost of ownership and eliminate time-consuming storage operations. It creates a cluster with a virtual switch and without external storage, where data storage is organized solely at the expense of server resources. The resulting computing power, disk space, and network components are managed from a single console.
Servers and special software for them can be purchased separately, but most often it is a ready-made solution from one manufacturer. Its advantage lies in comprehensive technical support and high chances of prompt elimination of problems that arise during the operation of the system.
And here we have approached one of the most significant issues.
Do you need to buy your HCI equipment from a single vendor, or can you combine them while still providing a single management solution?
We will examine this issue and give a detailed reply in two parts.
Combining vendors is usually not an option
Individual HCI suppliers mostly have substantial and private intellectual assets embedded in their particular software elements, and each option upholds a separate hypervisor.
As a result, the configurations of the solutions vary greatly. Some of them are entirely software-based; others are almost entirely software, but it includes an accelerator card that operates with some of the most complicated tasks, providing complex deduplication as well as other facilities without being concerned about additional effort and cost of them. Also, vendors – not all but often – have hardware compatibility lists to create vendor lock-in.
Thus, the brief reply is that mixing and matching manufacturers is not entirely feasible. For using alternatives from various HCI manufacturers, each must be implemented as a separate storage facility.
Combination of distribution channels
Assume that you purchased four devices from identical producers but varying distributors: two from the first and two from the second. All will be okay within this circumstance.
For example, consider one specific use situation. A number of vendors offer some prepackaged particular types of appliances. In hypothesis, you might probably buy two nodes from Dell and two others from HP. It seems that there are no practical reasons why you couldn’t incorporate all four nodes in the same grouping. However, if the hardware in this integration is not particularly distinctive, the proprietary software is crucial. Even though mixing hardware from multiple providers is entirely feasible, it is not recommended.
What are the advantages of such a system?
To support the IT infrastructure of a business, several components are needed. And each of these components needs to be serviced, which leads to the maintenance of various kinds of specialists (one person — one type of component). When scaling a business, an additional burden appears on everyone: different components need to be purchased and the delivery of all parts agreed, then find a place for this in a new branch (it is not always there), then take up installation and configuration, as well as connection. It’s a lot of fuss.
The goal of hyperconvergent systems is to solve all these problems and to solve them effectively, so among the key advantages the following can be mentioned:
- Such a single system combines all the necessary components and connects them at the software level. And this means that you need to purchase and install one system. Otherwise, we are talking about purchasing and installing several different components and linking them together independently.
- Such a system is easy to install. Generally, there is no need for any specific knowledge during installation and management. But suppliers of individual components may make it harder from time to time by adding clever proprietary systems that need mastering. There is no standard here.
- The administration is greatly simplified because the system is one, rather than a number of different components. This is important: according to various surveys, it is infrastructure support that takes the lion’s share of the time of those responsible for the functioning of the IT infrastructure. In our case, automation appears, which simplifies the time spent on maintenance, and at the same time the whole system is much simpler: it does not have any infrastructure features, like, for example, separate servers. Thanks to hyperconvergent systems, it is possible to avoid inflating the staff of support specialists and direct finances in other directions.
- Higher performance because the flash drives used are located inside the server and give less latency when working.
The post Can You ‘Mix and Match’ Hyperconverged Appliances? appeared first on The Alike.
