Аналитика и комментарии
DPC Increase Powers
Construction and development of data processing centers that would meet the needs of users in terms of reliability and performance is quite a pressing issue worldwide. It is no accident that the global market of DPC demonstrates a considerable growth in recent years. An important driver of its development is the dramatic increase of the stored data volume.
So far, the leader of DPC development is the United States: in this country there are three of the largest data centers in the world. It's a wonder that only one state - California - holds more than 200 data centers. This number is more than the current number of all data processing centers in Russia. Of course, it is fair to note that the Russian segment of the data processing centers began to develop much later than in North America or Europe. Therefore, the domestic market of DPC, on the one hand, develops in accordance with the global trend and has a lag of a certain number of years, and on the other hand, its late development allows to take into account all the latest trends that emerge within this industry.
SHORTAGE OF RESOURCES
According to the survey of Uptime Institute (an organization that has a world authority in the field of DPC, provides training, research and consulting in the field of DPC) 36% of companies are worried that they will face the problem of facilities and premises shortage. 40% of the companies are planning to build new data centers, while 29% of the companies expect to rent additional colocation spaces (server placement in a data center), and 20% of the companies intend to pass IT to cloud providers. Similarly, according to Afcom (Association for Computer Operations Management), which interrogated 360 IT professionals and directors, 29% of companies are expanding their data centers or establishing new ones, 21% of the companies plan to do so during the next two years.
There is also the fear of capacity deficit in Russia. According to the survey of CNews Analytics 49% of respondents indicated that they faced the lack of resources. They plan for the next year to establish their own site or use the services of commercial DPC. The main consumers of commercial DPC are medium-sized enterprises. Experts suggest that colocation services can be attractive for large organizations that have started the development of data centers quite a long time ago, so in the near future they are likely to face the lack of their own capacities or the need to modernize the DPC.
OWN OR RENTED DPC
Choosing between the construction of its own data processing center or the rent of data-centers, it should be remembered that both ways have their pros and cons. It is well known that large organizations prefer to build data processing centers, medium and small-sized companies have a tendency to rent such centers.
"In the first case the operating costs are saved, in the second case the capital costs are saved, says the director of Information Technology Department of Rosgosstrakh Bank Andrey Bondarev. -The availability of its own data processing center may improve capitalization, provide flexibility of its use". ln addition, your own data processing center allows you to easily implement various regulatory requirements of regulators and international payment systems.
According to Andrey Bondarev, rent of data processing centers may be divided into several options: space rent, rack space rent, computing powers and system storage space rent. "The first option allows the use of existing equipment and allows greater flexibility for its placement, including the bypasses of regulatory prohibitions, which is especially important for financial institutions. Rack space rent is beneficial only when you want to place a small amount of equipment, and there is no statutory prohibition on its placement. Finally, the last option is good for a start-up project, when you have to start with no extra costs for the purchase of equipment, its maintenance and administration. But in such case the operating costs are the most significant. Sometimes this option is used in the financial sector, but this is rather the exception than the rule."
CLOUD TECHNOLOGIES AND DPC
If we talk about the technological trends in the field of DPC, it may be noted that many organizations seriously consider and some of them actively use cloud technology and virtualization of servers and storage systems, knowing that this approach increases the efficiency of server platforms and storage systems.
As the volume of information increases and requires active work with all the layers of corporate data, DPC will also become powerful. According to experts, data storages, computer subsystem, network and engineering infrastructure - all the key elements of the data center will be developed by sequential increase of capacities. The performance and safety level will be improved, special attention will be given to business continuity and the recovery of activity.
Virtualization is one of load balancing means for resources utilization increase. It means running multiple virtual software environments or creating on the same physical device several virtual ones. Virtualization allows the flexibility of resources relocation needed to carry out business tasks with different performance requirements for hardware and software platforms.
"Virtualization had occupied a key place in the solutions of data processing centers, as it allows to manage data center resources with flexibility and use them more efficiently," - says the director of Operations Division of OTP Bank Maxim Pustovoy. Currently, DPC virtualization is increasingly mentioned in the context of the transition to cloud technologies. This is the main trend in the field of data processing center. "We gradually move to such solutions as IaaS (Infrastructure as a Service),"- the expert says.
Indeed, amid the increasing interest in cloud services commercial data centers adapt their business models to the new requirements of corporate customers. Moreover, DPC transition scenarios from the provision of simple services in respect of space rent for IT equipment placement (colocation) to the services of IaaS, SaaS and PaaS may be very different. For example, most companies launch cloud services, gradually expanding their range.
CHOICE OF SUPPLIER
When an external service provider of DPC is chosen, most of the respondents from CNews Analytics companies are primarily guided by the relevant level of Tier (89,5%). This is quite logical, because in the data processing center, which is certified by the level 4 Tier through the redundancy of all systems and the monitoring of their condition a high uptime is provided: if a failure occurs, the data processing center automatically reacts to it in a few minutes. The next criterion of provider selection is the availability of the data processing center outside of Russia (75%). An important point is the cost of services (39%).
Depending on the spheres the importance of these parameters slightly changes: the issue of price is more important for the energy (60%) and distribution (62%) sector rather than for banks and insurance companies (23%). Among other key criteria respondents noted, in particular, the issues of security and protection of personal data, "price - quality" and "brand - reputation" ratio of the supplier, including the reputation of its customers.
According to Maxim Pustovoy (OTP Bank), the choice of the DPC service provider mainly depends on the expected level of service: "for example, it depends on the following fact: whether you operate the equipment by yourself or expect that the operation will be performed by the DPC provider (in the case of IaaS type solutions). Generally speaking, the supplier must be certified in accordance with TIA standard, and the supplier must have a decent list of customers and the company status."
EXPERT OPINION
Applications, databases, portals, and other useful pieces of software are getting increasingly abundant in corporate server farms. IT departments have successfully virtualized enterprise-wide services such as e-mail and telephony, but next steps to total virtualization brought up some unexpected issues.
Let's look at a large organization consisting of multiple units running semi-independent projects, such as a city government or a bank. The "old school" approach was to give each unit a dedicated set of servers for its tasks. Instead, with advent of virtualization, one finds much reason in bringing all projects to a common corporate cloud thus expecting to slow down the hardware sprawl, to increase reliability, and, finally, to assure the information security.
However, having tried this in real life, IT department has found that, although creating a virtual machine is an easy and fast doing, the manual setup of a complex virtual computing environment consisting of server, storage, networking, and security components is not quite the same. When such orders start pouring down in a variety of volatile specifications with deadlines as tight as "immediately", their processing is getting increasingly error-prone and unsatisfactory. IT department faces a challenge of either hiring more engineers or failing to provide quality of service similar to "my server is my castle" paradigm.
Cloud automation is the solution to the problem. This is a software which sits on top of a cloud and gives business user the opportunity to create, modify and delete their computing environments automatically, eliminating the IT department as an intermediary between the user and their application. By ticking choices in a self-service portal, user selects number of virtual machines, types of operating systems, amount of memory and storage, configuration of a network, and applications to be installed. The order is then submitted to an orchestrator which selects an appropriate cloud, sets up and configures all required components of the virtual environment, and installs the requested applications.
Consequently the user gets back the agility they had had with dedicated servers; the IT department saves on an overqualified helpdesk; and the company enjoys a cloud which is economic, reliable and secure, and, more importantly - it is efficient.
