System scalability

Scaling possibilities

The platform can scale from a handful of telemetry devices to a nationwide deployment, supported by the architectural choices and techniques covered here.

The properties and characteristics of the IIoT Platform make it possible to scale the data collection system from single telemetry devices to tens of millions of units. The software can be deployed on users’ physical servers (resource-supplying organizations and others) or on cloud resources in dedicated, closed loops, with channel encryption provided by mobile service providers. External access to information is protected both by encryption tools and by firewalls and routers.

The platform can be installed on Linux/Unix-based operating systems. This lets the consumer ensure high fault tolerance and security, as well as maximum scalability. Building on such an operating system as a protection measure lets the consumer certify their information system at minimal cost while ensuring the security required by regulators.

Platform architecture

The platform architecture is based on:

• Application server — Ruby on Rails
• PostgreSQL DBMS
• Virtualization environment — KVM. The customer can install an application for virtual machine management on the servers and thereby gain high availability, fault tolerance, manageability of the solution, and greater scaling potential
• Web browsers — Chrome, Mozilla, Firefox, Safari, and others act as a universal client: information consumer and control panel
• The data collection server, which is based on the multi-threaded mechanisms of the GO platform
• A solution oriented toward the built-in security mechanisms of the Linux/Unix operating system
• Incoming data packets distributed by the load balancing mechanism; data transmission from the devices connected to the collection server is optimally configured by a special communication-session-spacing algorithm that avoids overloads

Telemetry modules connected to the data collection server operate in a single IP address subnetwork on GPRS and 3G networks, and on NB-IoT networks via the UDP protocol.

Reading and processing of data from metering devices takes place on the data collection server, and new types of devices can be added to the support server. The communication session time of a telemetry module is determined by the schedule set on the server, and that schedule can be updated remotely and dynamically through the web interface.

Maximum scalability of the data collection server is achieved in several ways:

• Database clustering by means of the PostgreSQL DBMS
• PostgreSQL DBMS backups taken on a defined schedule
• Virtualization, which makes it possible to apply fault tolerance, high availability, and similar techniques
• Additional stability and performance, provided by the complete absence of third-party libraries and modules (100% independence)

The system has low resource requirements for physical data processing nodes. For example, the configuration of a physical server to serve 100 000 devices requires 12 GB of RAM, 4 processor cores and 100 GB of disk space.