Converged and hyperconverged infrastructure
Two approaches are used to build IT infrastructure: convergent (CI) and hyperconvergent (HCI). In this article we discuss their differences, examine their pros and cons and show how hyperconvergence can be implemented using vStack as an example.
What is CI
CI stands for Converged Infrastructure.
In such infrastructure each element has dedicated roles (e.g. storage), which adds complexity and discreteness to the infrastructure. Such infrastructure requires a staff of specialists to maintain it, and increasing discreteness over time inevitably leads to an unfortunate result: infrastructure changes are slow.
What is HCI
Hyper-converged infrastructure, or Hyper-converged infrastructure (HCI), is
An IT infrastructure in which software-defined compute (SDC), storage (SDS) and network implementation (SDN) modules are combined into a single platform.
The fundamental difference with CI is that HCI has no dedicated role components. In essence, HCI is an infrastructure of identical blocks, each of which can perform all functions and roles at once.
Comparing converged and hyperconverged infrastructure with traditional infrastructure
| CI | HCI |
| Infrastructure can be complicated. | No overcomplicated infrastructure. |
| High maintenance costs. | High maintenance costs. |
| Functions are assigned to elements with assigned roles for each function. | All functions (SDC, SDS, and SDN) are assigned to servers whose roles are not allocated and are defined by code. |
Converged and hyperconverged infrastructure essentially perform the same tasks. The only questions are topology, cost, scalability and ease of operation.
An important factor is the inevitable complication of convergent infrastructure over time. The effect of such complication can be compared to a time bomb placed under the foundation of a building.
The vStack solution in the market of hyperconverged platforms
What makes vStack unique:
- The technology stack is “not like everyone else”: not Linux, not KVM, not ceph.
- Most of the solution architecture was created within the project rather than borrowed from an off-the-shelf opensource solution. This allows you to decide for yourself when and how the product will be developed.
- In vStack, we implemented all the best of the working solutions and took into account the bugs in them. This helped avoid mistakes that cannot be avoided at the product maturity stage.
- No legacy support. This results in good performance results and low virtualization overhead of 2-5%. Other platforms have 10-15% overhead. This is very important because Moore’s law is no longer working, and consumers are “fighting for every megahertz”.
What the consumer gets with vStack:
- freedom of choice of equipment;
- high economic efficiency;
- high performance;
- low virtualization overhead;
- vStack can be maintained by a single system administrator;
- economy on hardware;
- single control panel.
Who is vStack suitable for:
- service providers;
- development companies;
- educational institutions;
- healthcare;
- retail;
- industrial companies;
- fintech, banks;
- streaming services;
- GameDev;
- SaaS vendors;
- E-commerce;
- government agencies.
The vStack solution allows the product to evolve along the path that business owners need, rather than following the path of the programming community. The solution is built with the shortcomings inherent in mature enterprise-level products in mind, and produces good performance results.