7 problems that hyperconvergence solves
Twenty years ago, companies created and maintained their own IT infrastructure: purchased and configured data center equipment, upgraded and replaced outdated solutions. This approach consumed a lot of time and resources. That’s why hardware and software systems were developed to help optimize the work with data centers and create conditions for corporate IT development. Initially, these solutions were converged, but over time, hyperconverged infrastructure emerged. This helped solve business problems even faster and at a lower cost.
In this article we will tell you what problems of creating and supporting modern IT-infrastructure can be solved with the help of hyperconverged platforms.
Converged and hyperconverged infrastructure
First, let’s talk about the differences between converged and hyperconverged environments. Converged infrastructure or CI consists of several components with dedicated roles: server, storage, networking equipment and automation solution.
Soon a hyperconverged infrastructure or HCI was created. It combines software-defined compute, storage and networking layers into the same elements. Each node in the cluster provides all the layers. In hyperconvergence, there are no components with dedicated roles, and the discreteness of the IT infrastructure is replaced by the uniformity and monolithic nature of its components. This distinguishes HCI from CI solutions.
Companies face a number of challenges when creating and developing IT infrastructure. First of all, it is necessary to allocate a budget for expensive equipment, assemble a team and promptly make changes. Modern companies use both converged and hyperconverged solutions. Recently, however, there has been a trend towards HCI environments.
How hyperconvergence solves virtualization problems
In a converged approach, servers, network equipment, storage, SAN, NAS have separate interfaces, concepts, APIs, etc. When the IT infrastructure consists of many elements that differ from each other, it leads to discreteness, complicates operation and requires the involvement of narrow specialists. And discreteness tends to grow quickly. Making any changes to the infrastructure begins to take more and more time.
Discreteness of a hyperconverged system is much lower, and its elements are more homogeneous. Any changes to the IT infrastructure are faster and require fewer resources. For example, if one element (a server) fails, the company can simply move the disks from the failed server to a backup server and continue operating with the same level of redundancy.
The market for hyperconverged solutions in Russia is relatively young. There are two types of platforms: “full-fledged” and integrated. In the first case, there is no reference to a specific equipment manufacturer.
Integrated solutions can only be used with hardware from vendors, while in the case of “full-fledged” hyperconvergence, the customer chooses what hardware to use.
The latter includes the foreign platforms Nutanix HCI and VMware HCI. At the moment they have officially left the Russian market. Of the domestic solutions, full-fledged hyperconvergence is represented by vStack platform.
The Russian HCI solution vStack was developed by specialists of the venture startup vStack.com, which belongs to ITGLOBAL.COM LABS. The solution consists of software-defined data center elements. The vStack platform is included in the unified register of Russian software. As of today, vStack is practically the only full-fledged alternative to VMware, the virtualization market leader that left Russia.
What virtualization problems can be solved by the vStack hyperconverged platform
Problem 1. High price of equipment
Foreign virtualization market leaders offer the platform either together with embedded solutions or with a rigid HCL. For example, Cisco works only on its own hardware, while VMware works on partner companies’ hardware that has undergone long-term validation. Typically, this hardware is enterprise-grade. In addition to the high price it has a long delivery time. It is also worth noting that operating such hardware without vendor support often does not make much sense, because it often has “closed” or proprietary interaction interfaces.
The vStack platform is not tied to hardware and customers can choose their own vendors and use standard or commodity servers for certain tasks. Commodity is a versatile and easily replaceable piece of hardware, sometimes also referred to as “consumable” hardware.
Problem 2: Low performance
Often companies have to choose – save money and lose performance or increase budget. Reducing IT infrastructure costs with vStack occurs without sacrificing performance. On the contrary, vStack performance is higher than competitors due to the absence of legacy and irrelevant problems in the code base, as well as the lightweight nature of the technical solutions used.
Problem 3. High Overhead in Computing
CPU Overhead is an indicator that shows the decrease in performance of a virtual server compared to a physical server due to virtualization overhead. It tells you how much of the physical computing power is greater than the numbers required for the task at hand. The lower the value, the less money is spent on overhead.
The importance of this component is due to the fact that Moore’s Law has ceased to work. Technological support for CPU performance growth has become impractical. Where the consumer is used to solving such problems with more productive hardware, this is becoming impossible, forcing many to look for other ways to solve the problem. One such way is to reduce overhead.
The average CPU Overhead for many companies is 10-15%. In vStack this figure is 2-5%. Low Overhead allows the company to efficiently use the available capacity, reduce overhead and license costs.
Problem 4: Low efficiency under CPU Overcommit
CPU Overcommit indicates the ability to use more resources than are available. The higher the overcommit, the more cost-effective virtualization is. Its increase is due to the adaptive mechanism of work in conditions of CPU overcommit using several classes of resource budgeting. This allows vStack to achieve 900% efficiency in the CPU overcommit computation layer.
Problem 5: Shortage of personnel and high cost of specialists
To create, maintain and update the IT infrastructure in a company, you need a whole staff of experienced IT professionals. This is a minimum of one employee to support each server, network and storage system, plus professionals in the field of virtualization and IS. It takes time and budget to find such employees.
Problem 6. Scaling Difficulties
The more complex an IT infrastructure is, the more time and effort it takes to scale. In hyperconvergence, one node fulfills all roles, and all nodes are the same. This greatly simplifies the process of infrastructure expansion.
Problem 7. Large number of legacy
Many well-known virtualization solutions (such as VMware) have been under development for a long time. During this time, they have managed to accumulate legacy, that is, code that was created to solve problems that have already lost relevance. Its problem is that due to the constant increase of the number of layers, even a small extension leads to a loss of performance. And the user has to pay for the functionality he doesn’t need and is irrelevant. At the same time, removing legacy code is a long and costly process, and developers and owners are reluctant to do it.
There is no legacy in the code of the vStack platform. It is a modern, comprehensive, monolithic solution that appeared four years ago. The vStack developers do not address legacy issues and irrelevant problems, so that the platform remains flexible and lightweight.
Modern hyperconverged platforms help reduce IT infrastructure costs, increase productivity and accelerate the expansion and modernization of existing systems. At the same time, the cost of ownership of the Russian vStack solution is half that of VMware or Nutanix. This allows to solve problems and tasks faced by the Russian business in the period of turbulence without excessive costs.