Selectel Cloud Platform is a user-managed public cloud hosted in one of the availability zones. Each new user project is a separate isolated cloud. Cloud resources can be scaled.
Within a project, you can use the platform’s services to create an infrastructure of any complexity:
- Cloud servers, disk devices, networks;
- Managed Kubernetes;
- Load balancers;
- Serverless platform;
- Cloud storage.
|Project||A logically isolated resource group that a user can access|
|Quota (custom)||Limit of resources that can be used to create objects within a project. Quota is set when creating a project and can be scaled later on|
|User||The username and password pair created within a domain|
|Cloud server (Virtual machine)||Virtualized server created within dedicated resources|
|Account limit||Limit of resources that can be used to create objects within one account|
|Resource||The smallest possible unit of a service or server. When configuring your cloud, you can only choose necessary resources|
The following resources are available:
- Processor Cores (vCPU). Virtual processors built on:
- 2.3 GHz Intel® Xeon® E5-2670 v3 (Haswell);
- 2.4 GHz Intel® Xeon® E5-2680 v4 (Broadwell);
- Intel® Xeon® Gold 5218 (Cascade Lake);
- Intel® Xeon® Gold 6140 (Skylake);
- Intel® Xeon® Gold 6240 (Cascade Lake).
- Local disk. Solution based on NVMe SSD, more. Best for projects where read and write speed indicators are important.
- Network disk. SSD-based solution, more. Best for scaling disk space without changing the basic disk. Disk performance may vary depending on the selected Availability zone (see our website for details).
- Public subnet of any size from /29 to /24. Public («white») IP addresses are the best solution for high-load projects. Public addresses are used directly in virtual machines, as opposed to floating addresses.
- A private subnet is used to ensure reliable data transfer and to save on administrative traffic, both when combining cloud servers from different regions into one private network and also when combining dedicated and cloud servers into one private network.
- Floating IP address. A public address that can be reassigned to different machines. A solution for developing, configuring, and testing projects. Floating addresses are attached to a virtual machine’s private address and can be reassigned or detached without restarting the machine. Floating addresses are legitimate external IP addresses that are available until they are deleted by the client.
- Licenses. Separate licenses are applied for each server.
- User image storage. Disk space necessary for storing custom images. Storage is billed separately from standard disks.
- External traffic. Incoming and outgoing traffic from your infrastructure to the Internet. Traffic between services or within the Selectel network is not taken into account. All projects are provided with 3 TB of free external traffic every month. If you exceed this volume, the traffic is paid for separately. Please see detailed information on our website.
- VRRP subnet. This allows you to switch from the primary site to the backup site if necessary. VRRP (Virtual Router Redundancy Protocol) allows you to quickly switch a project from the primary site to the backup server (with which the data is synchronized). A database replica is being written.
- Graphics Processing Unit (GPU). Servers with Dedicated GPU cards are optimized for GPU computing, for example, for video transcoding, neural network training or creating remote workstations.
- LAN between the regions. There are two types of LAN: between the regions of different cities (based on a leased channel between data centers in St. Petersburg and Moscow) and between the regions of one city (based on its own channel between data centers in one city). Traffic volume is not taken into account in either case. The network is only limited by the channel width.
- Load balancers. Load balancers are used to distribute incoming traffic between existing cloud servers (virtual machines) within the same region.
Learn more about regions, pools, and availability zones.
Tasks to be Solved
Owing to flexible infrastructure configuration, Selectel Cloud Platform can be used in a wide range of tasks: from simple websites to complex scalable systems.
Placing internal systems on a private LAN will provide data and application security. The ability to use resources in different pools allows you to reserve critical systems and quickly add computing power with increasing loads.
You can use Selectel cloud platform in the following systems:
- A simple website or blog. High reliability combined with the ability to assemble a configuration that is best for your tasks allows you to provide a high-quality service without high financial and time expenses.
- Comprehensive web applications. Organizing the system structure in the cloud allows you to separate public web servers from applications that run business logic in a private local subnet. In addition, the simplicity and high speed of allocating additional resources allow you to create efficient services with a variable load without overpaying for unused capacity.
- Moving or expanding an existing system to a private cloud. If you need to quickly increase system capacity or create redundancy mechanisms, it allows you to configure a secured connection to the existing infrastructure and quickly complete your tasks.
- Information technology systems of the company. The flexibility of settings allows you to implement unique server configurations without worrying about system hardware.
Virtual servers containing information requiring high security and availability levels can be combined into local networks. To connect to the Internet, you can use a NAT connection on virtual servers outside secure private LANs in the cloud.
User accounts can be added to control developers and system engineers access to different projects. You can assign users to new projects, add SSH keys, and control access to the cloud in the personal account in the Control panel.
OpenStack Component Versions Used
The table contains a list of OpenStack components with links to API descriptions and the API version used.
|Component||Release version||API version|
|Nova||Stein||2.1 (maximum microversion — 2.72)|
|Cinder||Train||3.0 (maximum microversion — 3.59)|