CLusters

Серверный кластер

SOLTI designs, configures and maintains server systems: both corporate servers, originally intended for individual work, and groups of servers working together and performing common tasks.

The advent of the Internet has allowed businesses to master new directions. Trading platforms, remote banking services, personal accounts of users of government departments, electronic records in medical institutions make us think about the smooth operation and round-the-clock availability of Internet services. One computer is not able to cope with such a load. Failures and malfunctions lead to downtime, cause resentment of customers and result in financial losses.

To solve this problem, SOLTI suggests using clustering technology (server cluster organization). The cluster cannot be purchased in the finished version. Each of them is an individually customizable solution.

A cluster is a group of servers (nodes), combined into a single hardware-software resource with centralized management using high-speed communication lines to solve joint problems.

The cluster provides a level of reliability and high performance that is impossible when using a single server. If one of the servers fails, the tasks that it performed continues to be performed by the other server and the health of the whole system is restored. Due to the specific organization of work in the cluster, the system is perceived by the user as a single resource and the temporary loss of operability of the nodes is invisible.

The reliability provided by the cluster reaches 99.999%, and downtime can be reduced to 5 minutes per year.

Types of clusters

Fault tolerant

In a High-Availability cluster (HA), high availability is achieved by eliminating possible points of failure.

It is used in online trading platforms, continuous production, banking and exchange sectors

Benefits:

  • high availability of services and applications
  • reliability
  • downtime reduction

Balancing

In the load balancing cluster (NLB), the network load balancing service, which is installed on each server, allows achieving high performance. A message about a node failure (for example, a cluster of file servers or a cluster of database servers) starts the process of reconfiguring the system.

It is used in billing services of mobile operators, web hosting.

Benefits:

  • scalability
  • inexpensive nodes of standard architecture

Computational

The High-Performance Computing cluster (HPC) combines modules with powerful processors, high-speed communications and special-purpose software. This configuration is designed to perform demanding tasks in the industrial complex and research activities.

It is used to analyze Big Data data, study the potential of neural networks and artificial intelligence.

Benefits:

  • high performance
  • the possibility of parallel computing
Cluster Building Technologies

ACTIVE-ACTIVE. Nodes of equal priorities are involved. In the event of a failure, the operating functions distribute among themselves the remaining active nodes. After troubleshooting, automatic resumption of the original operation scheme is provided.

Disadvantage: the need to use special application cluster software

Advantage: increase the speed of the task.

ACTIVE-PASSIVE. It is universal for partial operation of working capacities. The passive module operates offline. The connection occurs with the received main server failure signal. When eliminating a defect, the administrator returns the system to its original state

Disadvantage: the impossibility of increasing the speed of tasks.

Advantage: cluster implementation is possible due to built-in OS mechanisms

Implementation example
Харьковский городской совет

Building a fail-safe cluster of virtual machines for the Kharkiv City Council

The customer set a task to design, build and configure a failover cluster of virtual machines to organize a corporate “cloud” infrastructure of the mapping system.

A two-node cluster based on HP servers was built. A virtual environment is deployed using Microsoft Hyper-V in the cluster. An information structure for the correct integration of the geoinformation system has been created. An external data storage system and a managed switch of the third level are installed. Management tools for servers, storage and UPS were configurated.

Details

CONSULTATION NEEDED?

Call us on +38 (057) 765-84-00 or leave your number / e-mail – we will contact you



Сертификаты