Building a data center

Consolidation and Virtualization

Popularizing services provided in the form of the so-called Clouds have changed the approach to designing data centers in recent years. Companies that, for various reasons, do not decide to maintain their own server infrastructure in external data centers, implement solutions that are their counterparts on a smaller scale - the so-called private clouds. Server rooms consisting of a large number of hardware servers, each of which provides a specific service and is a kind of "island" from the point of view of the network, disk system, hardware resources and management are a thing of the past. Modern server rooms constitute one complex, but flexible and easy-to-manage organism, the resources of which can be increased according to the growing needs of the business without unnecessary technological breaks. This was mainly due to the development of virtualization technologies (VMware, HyperV, Citrix) and mass storage (arrays, SAN networks).

Virtualization of server systems and the use of arrays as shared disk resources allows to eliminate a number of problems that have been faced for many years in data centers based on physical servers - "islands". The main advantages of virtualization and storage consolidation include:

Maximum use of hardware resources and incurred costs: 
In the traditional approach, the purchase of each new server was associated with a large financial outlay, which was used only to a small extent. The load of a typical server providing one service did not exceed 10-20%, which meant that a large part of the costs incurred did not bring any effect from the business point of view. Thanks to virtualization, it is possible to run from a few to a dozen or so virtual servers on one physical server, the resources of which can be used at the maximum level, leaving only a minimum reserve in the event of an unplanned load increase.

Eliminate downtime caused by failure:
The previously used servers in traditional terms - hardware + operating system - were a big problem from the point of view of restoring their functionality after a failure. A hardware failure was often associated with the need to install the system and services from scratch on new hardware, which often differed in parameters and required other drivers and configurations. The related downtime, even if backup servers or spare parts were kept, was unacceptable from the point of view of business continuity. HA (High Availability) solutions available in virtualization systems allowed for unnoticeable or minimal service interruptions to end users, even in the event of a failure of a significant part of hardware resources.

Increasing the performance and security of disk systems:
Disk arrays are a much more efficient and safer alternative to local disk systems in servers. The traditional disk system available in the server protects us only against failure of the disk protected by the RAID mechanism. The controller itself, its driver, cables, and motherboard are single points of failure, the replacement of which is associated with downtime in the server's operation. In addition, each local disk system requires some redundancy related to RAID protection and spares. By multiplying these redundant disks by the number of servers, we get quite a large pool of disks that we are not able to use effectively, and for which we had to pay specific money. Also, the performance of local disk systems in servers is a rather constant value that we have no influence on. We can eliminate all these inconveniences by consolidating resources on disk arrays. They offer much higher performance that can be scaled by increasing the number of disks. At the same time, the disks are used in a more efficient way, as the partitions shared between different servers share common redundant disks for RAID and spare. The reliability is also incomparably greater thanks to the redundancy at the level of disk ports, controllers, ports in the controller, cabling and ports in the server's HBA cards. Failure of any or even several of these elements does not have to mean an interruption in access to disk resources.

Reducing the disk space used and increasing the backup efficiency:
In environments where the amount of data stored can be problematic and difficult to maintain, it is possible to use storage systems that offer deduplication functions. Thanks to its use, it is possible to reduce the used disk space by several dozen to several hundred percent. It is very useful, especially when such large amounts of data have to be archived, backed up and stored for a long period.

Infrastructure management

The structure and dependencies present in a modern IT environment mean that the number of tasks facing administrators generally exceeds the capabilities of the IT team. Dozens of physical and virtual servers, network devices, mass storage, backup management, dozens of applications are the everyday work environment of people responsible for the IT system. IT infrastructure is one of the important elements enabling the implementation of basic business processes, while infrastructure failures threaten the business continuity of the enterprise.

The ITIL library (Information Technology Infrastructure Library) is a collection of good practices in IT systems management recognized as a standard in many companies. In the version 3 published in 2007, ITIL consists of 5 departments:

  • Strategy,
  • Design,
  • Transition,
  • Operation,
  • Continual improvement.

As part of the above studies, the following processes are carried out:

  • SLM - service level management,
  • SKMS - service knowledge management,
  • ISM - information security management,
  • CMS / CMDB - configuration management system.

As you can see, ensuring compliance with ITIL standards requires compliance with procedures and keeping appropriate documentation. Therefore, it is advisable that IT management support tools also comply with ITIL principles. Compliance with the standard increases the quality of user service, supports the continuity of the company's work and allows for a subsequent analysis of the functioning of the IT department.

Regardless of the support for ITIL, it is also important that the used software integrates well with modern complex IT systems in which physical and virtual resources coexist, there are many operating systems, local and wide networks, distributed datacenter resources, and many others.

As part of our cooperation with clients, we implement various IT infrastructure management systems - from free open-source solutions (eg OTRS) to chargeable management systems from Cisco, Microsoft, vmware. It is worth adding that in many cases, as part of the implementation of our service contracts, we use previously implemented systems.