PaaS: Basic Concept and Terminology

This guide introduces the core platform concepts used for application hosting across public, private, hybrid, and multi-cloud environments. It is intended for users who are new to PaaS and need a clear view of how the main platform components relate to one another.

The platform is built from several connected components: Cloudlets, Containers, Layers, Environments, Applications, Hosts, Environment Regions, Platform Installations, and the Cloud Union.

On this page

Cloudlet
Container
Layer
Environment
Application
Host
Environment Region
Platform Installation
Cloud Union
What’s next?

Cloudlet

A Cloudlet is the smallest platform resource unit. One Cloudlet represents 128 MiB of RAM and 400 MHz of CPU. This fine-grained allocation lets the platform assign resources close to the actual demand of each instance and supports usage-based charging.

Reserved CloudletsA fixed amount of capacity reserved in advance and charged regardless of actual use. Best suited to stable, predictable workloads.

Dynamic CloudletsCapacity added and removed automatically as demand changes. Best suited to variable or unpredictable workloads.

Reserved and dynamic Cloudlets — Reserved capacity forms the baseline, while dynamic capacity follows workload demand

Cloudlet calculation: The platform charges according to whichever resource is higher during the hour: RAM or CPU. For example, 2400 MHz of CPU equals 6 Cloudlets, while 1024 MiB of RAM equals 8 Cloudlets. In this case, usage is calculated as 8 Cloudlets, not 14.

Both reserved and dynamic usage may qualify for automatic volume discounts, so the unit price can decrease as consumption grows.

Container

A Container, also called a node, is an isolated virtualised runtime instance used for an application server, database, load balancer, storage service, or another software stack. Each container runs on a host and can scale vertically or horizontally.

The platform supplies certified containers for Java, PHP, Ruby, Node.js, Python, Go, and .NET, together with support for custom Docker containers. Every container has a private IP address and a unique DNS record, and a public IP can be attached through the dashboard or API.

Layer

A Layer, also known as a node group, is a collection of similar containers inside one environment. Common predefined layers include:

Load balancer (LB)
Compute (CP)
Database (DB)
Data storage (DS)
Cache
VPS
Build node
Extra custom layers for Docker-based environments

Layer of similar containers — A layer groups containers that perform the same function

Layer-level controls make it possible to restart, redeploy, configure, monitor, and horizontally scale all matching containers together. Scaling can be performed manually through the dashboard, API, or CLI, or automatically through load-based triggers.

Containers in the same layer are distributed across different hosts through anti-affinity rules, helping improve availability and resilience.

Environment

An Environment is a collection of isolated containers and layers used to run a specific application service or group of services. Built-in tools support configuration, monitoring, troubleshooting, and resource management.

An entire environment can be started, stopped, cloned, migrated to another region, shared with team members, and monitored for resource consumption.

Each environment receives an internal third-level domain automatically. A custom external domain can be connected through a CNAME or A record, and domains can be swapped between environments for traffic redirection.

Application

An Application is the complete project formed by one or more environments. A simple application with only one or two software stacks may fit inside a single environment, while a complex application is usually divided across multiple environments for easier independent management.

Tip: Separating application components into different environments gives greater flexibility during deployment, updates, scaling, and maintenance.

Application made of several environments — An application can combine several independently managed environments

Application source code or compiled packages can be deployed from:

A Git or SVN repository with automatic update support
A local archive such as ZIP, WAR, or EAR through the dashboard or FTP
A custom Docker image
A pre-packaged solution from the Marketplace

Host

A Host is a physical server or a large virtual machine running on technology such as KVM, ESXi, or Hyper-V.

Hosts are divided into isolated containers that form environments. This model supports multitenancy, high workload density, and more efficient resource use by distributing containers across available hosts according to demand.

Environment Region

An Environment Region is a group of hosts coordinated inside one isolated network.

Each region has its own datacentre capacity, public and private IP pools, and resource pricing. Projects can be migrated between available regions when a different location or capacity profile is required.

Platform Installation

A Platform Installation combines multiple environment regions with a central cluster orchestrator so that they operate as one system. Regions may be placed in separate datacentres or clouds, giving users different options for location, capacity, hardware, and pricing.

This architecture creates a multi-datacentre or multi-cloud environment that can be controlled from a single panel. Each installation is operated by a hosting provider and supported by its local team.

Cloud Union

Cloud Union is the global network of hosting providers that distribute and operate the platform. This model gives customers a choice of local providers, jurisdictions, support teams, datacentre locations, performance levels, and pricing models.

The partner ecosystem is designed to reduce vendor lock-in and allow customers to select a provider without compromising on location, support, performance, or cost.