Locations: countries, regions, availability zones, pools, and data center addresses
With Selectel's products and services, you can create an infrastructure that meets all requirements for fault tolerance and security.
Selectel infrastructure is located in various locations.
A location is the physical placement of your infrastructure within Selectel's infrastructure, which includes:
Choosing a location affects how you manage availability, fault tolerance, and load balancing in your infrastructure.
Currently, you can choose from 4 countries to host your infrastructure in Selectel, which are divided into 6 regions, 11 availability zones, and 26 pools.
You can check the availability of Selectel products and services in different locations in the availability matrices.
Country
Country (e.g., Russia) — this is one or more regions that are located within a single state.
In most cases, splitting by country does not affect anything, and the main infrastructure division occurs at the region level. In some cases, there are restrictions for regions in different countries:
- when using a global router, you cannot set up network connectivity between products located in different countries. You can configure connectivity between products and services only between different regions within the same country;
- in S3, different domains are used for each country, and TLS (SSL) certificates are bound to a country — the certificate only works for buckets hosted in a region within the certificate's country.
Region
Region (e.g., Saint Petersburg) — these are several availability zones with Selectel data centers and partner data centers located within a major city's metropolitan area.
Data centers within a single region are usually located relatively close to each other, within tens of kilometers. They are connected by a fiber-optic communication line with data redundancy, and minimal signal latency is ensured between data centers within the region. Learn more about data centers on the Data centers page on selectel.ru.
Data centers in different regions, by contrast, can be located hundreds or even thousands of kilometers apart. Each region is completely isolated from others: independent connection to power lines, autonomous power and cooling sources, and dedicated communication channels. This allows for geographical distribution of infrastructure and its fault tolerance.
Selectel provides connectivity between regions via redundant, high-performance communication channels.
To host your application, we recommend choosing the region closest to your users as the primary one. You can place resources in different regions — then the infrastructure will be able to survive potential provider-side failures with minimal downtime, for example, Internet outages at the border level, natural disasters, and emergencies.
You can view the list of regions in the Selectel Infrastructure table.
Availability zone
An Availability Zone (example: Availability Zone 1 in Moscow) is one or more data centers within a single region. Availability Zones within the same region can be located at different distances from each other—from a few meters to tens of kilometers. Each data center in an Availability Zone is equipped with autonomous power and cooling sources and its own redundant communication channel with ultra-low latency.
Selectel reserves all critical system elements within an availability zone: power supply, cooling, and network infrastructure. Each availability zone is provided with 24/7 physical security and monitoring. Equipment maintenance operations are not performed simultaneously in multiple availability zones within the same region.
We recommend that you reserve critical parts of your infrastructure in multiple availability zones. There is no single point of failure for availability zones—only a single point of failure for the entire region. This will protect your system in the event of a failure in one of the zones, such as a data center power outage, fire, or natural disaster.
You can view the list of availability zones in the Selectel Infrastructure table.
Pool
Each data center and, consequently, each Availability Zone contains several pools. Each pool is isolated from hardware and software failures occurring in other pools.
Network connectivity within a single-zone or multi-zone pool is ensured by an L2 connection. To configure connectivity between pools, you can use the Selectel global router.
For cloud infrastructure, pools are additionally divided into pool segments (example: ru-2a, ru-2b). Segments provide additional fault tolerance—infrastructure in different segments is placed in different racks within a single pool. The fault domain for a segment for servers and storage systems is an individual rack. Network infrastructure operates at the pool or Availability Zone level.
You can view the list of pools and segments in the Selectel Infrastructure table.
Single-zone pool
Single-zone pool (e.g., ru-9, SPB-4) — part of the infrastructure in one of the data centers of an availability zone.
Segments of single-zone pools are located in different racks within the same data center of an availability zone.
Placing infrastructure across multiple pools protects against network hardware and software failures (for example, network equipment failure or power circuit failure) and reduces the likelihood of data loss and service downtime. The single point of failure for single-zone pools is a failure in the Availability Zone.
Multi-zone pool
Multi-zone pool (e.g., ru-6) — part of the infrastructure in multiple data centers of different availability zones.
Segments of a multi-zone pool are located in several data centers in different availability zones.
Placing infrastructure in a multi-zone pool:
- protects against the effects of natural disasters and fires in data centers;
- protects against network hardware and software failures (for example, network equipment failure or power circuit failure);
- reduces the risk of data loss and service downtime.
A single point of failure for a multi-zone pool consists of failures in all availability zones where the pool segments are located.
Increase fault tolerance
You can increase infrastructure fault tolerance at the following levels:
- hosts;
- racks or pool segments;
- pools;
- availability zones;
- or regions.
For example, to improve fault tolerance of cloud servers at the host level, you can place cloud servers on different hosts in a placement group. To improve fault tolerance at the rack and pool segment level, you can move a dedicated server to a different rack, and a cloud server to a different pool segment. For fault tolerance at the Availability Zone level, you can place servers in different Availability Zones, and for cloud servers, use segments of the multi-zone pool ru-6.
Selectel Infrastructure
Russia
Uzbekistan
Kazakhstan
Kenya
St. Petersburg
Moscow
Novosibirsk
Data center addresses from the state address system (FIAS)
Exact cadastral addresses of Selectel data centers for contracts, official communications, letters, government requests, or internal documentation. Addresses on maps may differ from registered office addresses.
St. Petersburg
Moscow
Novosibirsk