Requirements of distributed store designs

Interconnection transparency

A distributed store system should hide the connection details from the user. They should not (need to) know any of the details concerned with the means of connectivity (ie: ethernet).

Location transparency

The location of the resource that a user is accessing should not be obvious to the user. Accessing a resource on a remote node should not be a explicitly different process from acting a resource located locally.

Naming transparency

The user should not be aware of the naming conventions used to implement the distributed store. Naming conventions for remote resources should be similar to those for local resources.


Each node in the network should have the same view of a shared resource. When the shared resource is altered in some way by a node all nodes in the network should be immediately aware of the change.


Only those with sufficient permissions to access a shared resource should be able to do so.

Fault tolerance

It should detect and recover from faults in the system.

High availability

The store should be accessible as much of the time as possible. This includes having requests acknowledged as quickly as possible after they are sent by a remote node and serviced without a waiting period. It also means when faults occur in the system they are isolated in the nodes they affect.

Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License