Queue job for the computation of the centroid of a cluster

Various helpers for DataAccessors

A default serializer, using the .NET binary serialization process. Requires TEntity to have the [Serializable] attribute. It cannot be used because of Windows azure security policy unless you have full trust rights.

A benchmark to evaluate the results of an implementation of k-mean. It is a lazy object which will return its analysis through its properties.

An abstract class that represents one of the entities used in the k-mean algorithm

The class used to store which entity is associated with which centroid.

Queue job for the computation of the cluster of an entity

Reader/writer of cloudster entities to/from the azure blob. As data must not be modified, no update methods are provided.

Data accessor for local (dictionary based) storage

A class used for logging in cloudster.

Queue job packing several other jobs.

Windows azure job.

A pure sequential (i.e. using a single role) implementation of the k-means algorithm

Configuration for cloudster : account informations, tables names, queue names, blob container names...

A wrapper for a table containing metadata used by all roles, and that must be shared.

Helpers for nicely using Stream objects These enable us not to use StreamReader's, which have some downside, such as closing the underlying stream when closed

Data accessor for Azure based storage

Wrapper for a class containing all the tasks enqueued and not completed.

Worker type.

Interface for a centroid computer. The method Compute computes the centroid of a set of entities.

Interface for accessing entities storage and bindings

Interface for a distance between two Entity objects.

Interface for an entity serializer, that translate data in a stream into a real entity and vice-versa. All members must be thread safe (if they are called on different streams).

Log level of a log message. They are the same as those accepted by WriteToLog(String, String)

A phase of execution.

Status of a given task.