BIS architecture diagram (zoom on the building scale)

At the building scale (intermediate scale of the overall system), the architecture of the Building Information System (BIS) is represented by the diagram below:

The detail of the elements of the heritage layer (above) and of the terrain or Edge layer (below) are not shown. The detail of the schematized building information system (BIS) includes the following elements:

• The SpinalCore BOS, the heart of the platform, which performs the functions of transverse management of system data and the management of the digital twin

  • An SDK allowing the development and deployment of services connected to the BOS and orchestrated by the BOS

  • A suite of services / micro-services divided into several categories:

    • The API gateway to third-party IT services of the legacy layer or the building layer

    • The API gateway (or driver gateway) to third-party OT systems in the field layer

    • Native building layer applications built directly on the basis of the SpinalCore SDK

    • Data sources for the building layer (BIM, DOE, Occupancy, room reservation services, agenda, etc.)

Software kernel (Kernel), parallel with the PC

A system kernel (or simply kernel, or kernel) is a fundamental part of the operating system.

Under BOS, a managed asset is not "just" a computer (a system) but a building that itself has subsystems (a system of systems). In other words, the BOS is a "virtual OS" that must communicate, orchestrate and manage data coming from other distributed systems in the building.

As part of the operating system, the kernel provides:

• Hardware abstraction mechanisms: sensors, actuators, controllers, structure, furniture, building equipment, complete building. The digital twin brings this abstraction.

• Information exchange mechanisms between software and hardware devices (via API or SDK). The kernel also enables various software abstractions and facilitates communication between processes and process orchestration.

The core of an operating system is itself software. As part of the BOS, this is the SpinalCore data hub having:

• Semantic graph management skills to manage the digital twin (or Building Graph)

• Real-time or asynchronous management skills to manage “hot” data

• Data replication and graph skills to distribute complex data and abstraction mechanisms to other systems ...

Its central role also requires high performance. The SpinalCore kernel therefore implements an “in-memory” database management system and is deployed on servers rich in RAM. This makes the kernel the best performing part of the operating system.

The diagram below shows the different parts of the core:

Functions performed by the nucleus

The core function of the kernel is to manage the digital twin of the building, to manage data flows between applications and to provide an interface between the kernel space and user space programs (SDK).

Besides the functions listed above, the kernel will also provide functions such as:

• File systems management

• The orchestration of several specialized schedulers (batch, real time, inputs / outputs, etc.)

Nature of BOS SpinalCore

The BOS SpinalCore is a data management platform based on a graph, asynchronous, in-memory and distributed database management system. There are several types of data management platform that the ONA generally classifies into three categories:

Data Lake: A “data lake” is a system or data repository used to store data in its natural format. Invented in 2010 by James Dixon, then CTO at Pentagon, this term generally refers to a single store of all corporate data in the form of files or blobs. Whether it's data from the source system or other unprocessed data, a data lake refers to a large pool of raw data for which the purpose has not yet been defined.

Example: MongoDB DBMS

Data Warehouse : A “data warehouse” is a data storage system used for reporting and analysing data. Also known as an enterprise data warehouse, this type of repository system processes data uploaded directly from a company's operational systems. Unlike a data lake, a data warehouse only processes structured data, which offers storage space and accessibility benefits to a larger audience.

Example : SGBD SQL

Data Hub : A Data Hub, or data hub, is a virtual storage platform. It brings together a set of data from multiple information systems that are homogenized to be redistributed.

The goal of the Data Hub is to provide businesses with a unified, centralized data source, so third-party systems can quickly and easily access the information they need.

Example : SAP Hana

→ The BOS SpinalCore is, according to the previous definitions, a data hub dedicated to the management of physical asset digital twins.

Physical or virtual server

It is recommended to deploy one server per BOS (physical or virtual machine) and to deploy on this server only the BOS and its native applications. This is to ensure network availability, the availability of RAM and more generally of all the resources necessary for the proper functioning of the BIS. The characteristics and expected performance depend on the building on which the BOS is deployed. It is possible, on large buildings, to consider the building as a heritage and to deploy several BOS and a POS.

The minimum skills of the servers to be provided will be as follows:

• A Linux OS (Ubuntu 18.04 LTS)

• 4 Gbit (Gigabit) of RAM minimum per managed building (to be adapted according to the size of the building)

• A 4-core processor and 2 GHz minimum

• Gigabit network card

RAM and computing power capacities may change depending on:

• The size and complexity of the building (number of GTB, SSI points, etc. to be ascertained)

• The number of applications deployed above the BOS and their nature

• The volume and history of data to be kept in the BOS in-memory database

• It will be necessary to dimension the server for each project.