Distributed data centers from the provider: what and why

Over the decades of its existence, data centers have come a long way from small computer rooms to computer centers, data centers, cloud and distributed data centers. What are distributed data centers for? The well-known Cisco Systems highlights the following goals for their creation:

• Disaster tolerance
• Data processing continuity
• Application mobility
• Systems migration
• Capacity / Capacity Building
• Distributed Services
• Geographically-localized services

Medium-sized organizations often have a fairly centralized model, with all data, services, and applications concentrated in one data center, which provides guaranteed levels of data availability and security. However, it is often necessary to have a backup data center integrated with the main data center using high-speed communication channels into a single computing complex.



The system of distributed data centers (the main data center and several regional), which perform the functions of collecting, storing, processing and delivering data to support corporate business processes, can be the best solution for large geographically-distributed organizations. Regional data centers can also be used as backup sites, use them at peak load, or implement a disaster recovery solution (DR). It will ensure business continuity even if a significant portion of IT resources or communication channels fail.

However, creating your own data center is a costly project. When calculating the cost of building a backup data center, it is necessary to take into account IT infrastructure costs and licenses / software support. By some estimates , this is more than 60% of the total cost of creating a data center.


Most often it makes sense to contact the data center service providers, and not to invest a lot of money in the construction and support of its own site. It is not by chance that in the recent difficult years in Russia , the number of projects for the transfer of IT infrastructure to outsourcing has increased on the part of large retail, social networks and some IT companies. Renting an IT infrastructure is more economical and gives additional business flexibility, allows you to launch new services and develop existing products without capital expenditures.

Usage scenarios

Distributed data centers minimize various risks, and the recent blocking of sites and subnets by Roskomnadzor, which in particular caused interruptions in the work of many cloud services, shows that any surprises are possible. Some companies see a way out of this situation in acquiring a server abroad, on which to organize their VPN, and to start all the necessary cloud services through it.

In the case of redundancy and some services, access speed to the site is critical, and providers take this into account. For example, RUVDS RUCLOUD's own data center in Korolev has three 5 Gb / s Internet channels, which gives guaranteed high speed. The channels are connected to the largest communication centers M-9 (MMTS-9) and M-10. Equipment customers can choose from any of the sites. Due to this, the user gets a speed of at least 250 Mbit / s with access in Russia, which allows him to work comfortably with his virtual server (VPS / VDS).

For example, the RUCLOUD data center is connected by a direct communication channel with a capacity of 10 Gbit / s with the main communication node in the data center M-9. This allows customers to guarantee a high rate of traffic exchange between virtual servers located in different data centers.

In addition, providers have the opportunity to improve the efficiency of backup data centers by maximizing the use of IT infrastructure resources (economies of scale not available in most corporate data centers).

In a typical distributed data center scheme, local area networks (LAN) and storage area networks (SAN) sites are interconnected. The local network segments to which the servers are connected are combined into L2 domains, and this allows applications to transfer IP addresses of servers between sites transparently. By combining SAN servers can use the storage resources of different sites.

From the point of view of management and operation, geographically distributed data centers (there are four types of such data centers) can look to the user as a single system. When a distributed data center provides services to external users, a single interface of services and support is available.


Classification of distributed data centers by distance between sites.

Distributed Data Center

What does this give?

What else do distributed data centers provide to the provider and its customers? For the provider company, they mean the possibility of expanding the scope of activities, entering new markets, allowing them to offer customers new points of presence and more choice. Geographically distributed international data centers can provide the best conditions for different parameters. In addition, such providers, as a rule, have a flexible pricing policy, which allows them to choose the best option for the placement of IT equipment and the tariff, taking into account individual requirements.

In particular, for the client it is:

1. Ability to create various disaster recovery configurations (Disaster Recovery) with preset disaster recovery settings. Backup data center can also be used to host IT infrastructure to ensure high availability of company applications. For example, the blocking by Roskomnadzor of “extra” IP addresses had a negative impact on many Internet services and sites that were not related to Telegram. In case of such blocking of IP addresses in one country, part of the infrastructure in another will come to the rescue.
2. The ability to organize work as close as possible to the desired target markets and to minimize network latency due to “bringing the service closer to users”. For example, Switzerland and England are often chosen by traders selling on European markets, owners of online stores.
3. Ability to work in several jurisdictions , observing local laws, for example, the requirement to process customer data in the territory of the Russian Federation, the EU, the UK
   
4. The possibility of simple migration to a new site with the preservation of data and even funds in the account. For example, if a client uses servers in the Russian Federation and decided to transfer data to London, then it is enough for him to submit a request to the support service, and after a short time his data will “move” to the British capital.
5. Distributed cloud data centers are also considered the best solution for providing IaaS services.

Disaster tolerance and high availability

As noted above, by placing IT systems in several data centers, their resiliency and high availability can be ensured. Each data center plays its own role, but in general, the independent implementation and support of such an architecture is usually not a trivial task. Moreover, the concept of high availability depends on the specifics of the company's business. In order to take advantage of such a topology with several data centers, we need an appropriate implementation of the data center architecture, the applications and processes functioning in them.


Catastrophe-resistant and high availability of IT systems can be provided by placing them in several data centers.

Disaster Resilient Data Center (DR) is used to host the IT infrastructure of business-critical applications to support IT services in the event of a disaster or a crash of the primary data center. In this case, it is necessary to ensure an acceptable level of losses during migration from the main data center in an emergency.

There may be several backup sites, depending on:

1. Critical systems for the main activities of the company.
2. The risks and potential losses from downtime IT systems.
3. Requirements for the performance and reliability of IT systems.
4. A selection of systems requiring redundancy and high availability.
5. The cost of organizing a backup data center.
6. Target data center topology.

Some functions of the backup data center can take on a public cloud. Placing applications in public clouds in order to increase flexibility and optimize IT infrastructure costs often becomes the most sensible solution. Therefore, in the long term, it is possible to reduce the number of data centers due to the use of public cloud services.

Catastrophe-resistant involves compliance with the required RPO / RTO-recovery point (Recovery Point Objective, RPO) and recovery time (Objective Restoration, RTO). The smaller the RPO / RTO, the more expensive the solution. In this case, data replication and “hot backup” of systems are commonly used. With a larger RPO / RTO, you can get by with backing up to tape or cloud and “cold backup”.

To ensure business continuity, geographically distributed failover clusters can be used. Cluster systems typically require L2 VLAN “stretching” between data centers, although some applications support clustering over the L3 network.

Virtual service mobility means the ability to move virtual machines between data centers. This requires stretching the VLAN and ensuring continuous access to the LUN volumes. Such a solution can be the basis for managing cloud load, balancing it and migrating the load between sites.

At the same time the distance between the sites is a key factor. With a small distance less latency, better performance, easier communication. You can use synchronous replication.

Hermetic zone in london

Providing a provider of disaster tolerance and high availability for critical client applications is one of the most sought-after scenarios. Another option is local backup. For example, the company RUVDS, which provides virtual server services (VPS / VDS), at the request of clients, organized the reservation of their virtual infrastructure inside the main data center, when client servers were duplicated on different physical machines, in different containment zones and at the same time connected to different network infrastructure. Then to this was added the ability to duplicate (or use as the main) server in Switzerland , and now in England.


With such a diversity of sites, of course, there is the problem of limited speed of data transmission between countries, but it is enough to organize additional reservation of critical resources. At the request of customers, mobility of virtual services is also possible (moving the VM between data centers).
Following the opening of the second site in Moscow, RUVDS launched a new pressure containment area in London - the financial capital of Europe. HERMOZONA is located in the Equinix LD8 data center - one of the largest data centers of the Equinix network. The LD8 data center has a long history and is considered one of the most important data centers outside the Telehouse campus. In 2016, Equinix bought Telecity for 2.6 billion pounds sterling, and both buildings were renamed Equinix LD8.


Equinix LD8 is a N + 1 redundant data center that fully complies with TIER3 requirements and OHSAS 18001, PCI-DSS, ISO 14001, ISO 27001, ISO 9001, ISO 50001, PCI-DSS specifications (the latter is important for payment systems).

Equinix serves customers from all over the world and 175 data centers in 22 countries. Equinix LD8 data center is located just a few kilometers from the London financial center, is connected directly to the online exchangers in Frankfurt (DE-CIX) and Amsterdam (AMS-IX) and is one of the best not only in the UK, but also in Europe .

In this data center, customers get access to various financial services, Internet service providers, cloud services and IT services, services for enterprises of various profiles, such as content developers and digital media companies.

The data center has an ISO 27001 information security certificate, an ISO 50001 power management certificate, an ISO 9000 quality standard certificate, an ISO 14000 environmental management standard certificate, an OHSAS 18001 industrial management certificate, and a PCI DSS card data security certificate. The entire infrastructure of the data center is reserved under the scheme N + 1.

The partnership agreement will allow RUVDS not only to expand operational opportunities and improve the quality of the services provided, but also to remain one of the most dynamically developing hosting providers in the entire post-Soviet space. A private zone in London will now be added to its own data center in Korolev, which is leased by RUVDS to the containment area in the Swiss data center Deltalis.
An isolated containment ensures that no one other than RUVDS customers will use the network infrastructure that is allocated to it. This option is the most appropriate in terms of price-quality-reliability indicators.

The hermetic zone is equipped with dual-circuit cooling (hot and cold corridors). It uses Huawei enterprise-class equipment. High quality of production, reliability, warranty service at a reasonable price allow us to provide our clients with rental services of dedicated virtual servers (VDS / VPS) on high-performance new equipment at an attractive cost.

Servers Huawei RH2288H V3 enterprise-class high density. These reliable and cost-effective server platforms help the provider to optimize the cost of services provided to customers.

Huawei is a strategic partner of the Russian hosting provider RUVDS since 2016. In particular, as part of a joint project, Huawei equips the data center RUVDS, located in the Moscow region , with its equipment. This guarantees a faster and more stable operation of virtual servers, allows us to offer our customers even better and more modern services.

Platform based on MMTS-9

RUVDS also opens a new platform on the basis of MMTS-9 ( M9 ). Currently, MMTS-9 is the largest technological platform in Moscow for interaction between Moscow, Russian and international telecom operators in the provision of local, long-distance, international telephone communication, data transmission, Internet access and other telecommunication services. In addition, MMTS-9 is the largest point of inter-operator exchange of Internet traffic in Russia.

At all three RUVDS sites, prices for customers are the same - from 130 rubles per month.


The building of the station MMTS-9 is equipped with modern engineering infrastructure, including uninterruptible power supply systems, climate control, gas fire extinguishing, monitoring and video surveillance. The presence of high-tech engineering infrastructure allows MMTS-9 to ensure the organization of a fault-tolerant point of presence with the necessary conditions for the functioning of the server and telecommunications equipment of telecom operators.