MLPACK is a C++ machine learning library with an emphasis on scalability, speed, and ease-of-use. Its aim is to make machine learning possible for novice users by means of a simple, consistent API, while simultaneously exploiting C++ language features to provide maximum performance and maximum flexibility for expert users.
ClodHopper is a Java library for high-performance clustering of numerical data. It contains clustering implementations such as K-Means, K-Means++, X-Means, G-Means, Fuzzy C-Means, Jarvis-Patrick, and various forms of hierarchical clustering. ClodHopper's clustering implementations take advantage of the host system's concurrent processing ability to speed clustering. The data structures are also very lean to conserve memory usage. ClodHopper is very extensible. If you are developing a new clustering algorithm, you may save yourself an enormous amount of work by extending a ClodHopper base class.
Hados stores files in a cluster of servers. Its goal is to handle high availability by storing copies of the same file on several nodes. It provides RESTFUL APIs to easily store, check, or retrieve files. Using the cluster APIs, you can retrieve files from whichever node hosts them. To avoid any single point of failure, it is possible to apply a request to any node of the cluster; there is no master node.
K-tree provides a scalable approach to clustering by combining the B+-tree and k-means algorithms. Clustering can be used to solve problems in signal processing, machine learning, and other contexts. It has recently been used to solve document clustering problems on the Wikipedia collection.
Proto Balance Mail is an enterprise SMTP cluster solution that supports distribution of email accounts. It scales up to 1,000,000 mailboxes apportioned over up to 125 backend mail servers (8000 mailboxes per server). No NFS or SAN is required. SOA is configurable with SOAP/XML. Anti-spam settings can be set per-user. Grey-listing is supported. Mal-ware is automatically detected and infected client PCs are automatically black-listed. POP load balancing is done. SMTP AUTH is supported. There is a Web-based management interface. Spam blocking is done by on-the-fly connection behavior analysis. It handles up to 10,000 concurrent SMTP connections. Streamlined CRM integration is done with HTTP+XML posts. Email-alias lists, forwarding, and out-of-office auto-reply are supported.
jmemcached is a fast network available cache daemon. It is protocol-compatible with memcached, but written in Java and suitable for applications with portability concerns, where Java is the preferred solution, or for using the memcached protocol in embedded applications with alternate storage engines. Existing clients for memcache work unmodified. It can run as a standalone daemon or be embedded inside an existing Java application.
BorderFlow implements a general-purpose graph clustering algorithm. It maximizes the inner to outer flow ratio from the border of each cluster to the rest of the graph. The main advantage of the algorithm is that it does not need parametrization to compute results of high accuracy.