Close-in computing is defined as: cloud computing that uses user-owned servers, equipped exclusively with free and open source software, located in rooms that the user controls as owner or tenant, the user having exclusive physical access and exclusive root access to the servers.
The fact that the servers are equipped exclusively with free and open source software means that all the programs installed on the servers, from the various applications to the operating system and to the firmware of any hardware component, are free and open source software. The rooms where the servers are located can be rooms in an office building or rooms in a residential building. Exclusive physical access and exclusive root access (for one individual or for a group of individuals who collectively own the servers and collectively own or rent the rooms in which they are located) is a condition for digital security and privacy, which are among the main goals of close-in computing.
Four critaria must be met by a server cluster to be considered a close-in computing server cluster:
- The servers must be owned by the user.
- The servers must be exclusively equipped with free and open source software.
- The rooms in which the servers are located must be controlled by the user, either as owner or tenant. The user can be a single individual, or part of a group of individuals who collectively own the servers and collectively own or rent the rooms in which the servers are located.
- The user must have exclusive physical access and exclusive root access to the servers, either as a single individual, or as part of a group of individuals who collectively own the servers and collectively own or rent the rooms in which the servers are located.
A room is defined as: an area within a building that has been set apart from surrounding space by standard building walls.
Various companies can provide close-in computing certification services. To certify that users have exclusive physical access to the servers, the certifying company can use tamper-proof security seals to seal the server cases, lock the servers in server racks and hand over the keys to these racks to the users, specifying this in a written document. They will also require the users to prove that they have the keys (proximity tags/access control cards, etc.) needed to access the servers from outside the building where the server rooms are located.
It should be emphasized that close-in computing provides users with the maximum possible level of control over their data in server-side computing. This results in a much higher level of security and privacy for this domain of computing.
To be noted that a close-in computing server cluster represents a close-in computing environment only for its owners who are also the owners/tenants of the rooms in which the servers are physically located, but not for other users, who may use the services hosted on the servers but don’t own the servers, don’t own or rent the rooms in which they are located and don’t have physical access or root access to the servers.
The close-in computing concept has a confrontational character because it was created as a reaction to the standard ‘cloud computing’ model, namely as an alternative that offers users the maximum possible level of control over their data for server-side computing.
Although this concept may seem far-fetched, a close-in computing server cluster is achievable today by using mainboards certified to be exclusively equipped with free and open source software by the Free Software Foundation, like the following mainboard, certified in 2019:
In a future article we’ll explain how to use the load balancer HAProxy to create a high availability server cluster from servers using mainboards like the one mentioned above and free and open source applications like those that make up RED SCARF Suite.
