An approach to XMPP server project

Recently, the company needs to build XMPP server, after the research has focused on the service plan between Tigase and Ejabberd, but with the depth of research has gradually found some of their differences, the first biggest difference lies in their cluster strategy, In the end, we chose Ejabberd, and there is no reason why his clustering strategy is free to support the expansion of water products.

The advantage of Tigase is that Java is written, expand its entry threshold is relatively low, complete data, but the free strategy of the cluster only supports the maximum of 3 too ha solution, and does not open the server cluster policy, need to use to buy advanced cluster jar, seemingly very expensive 2000$/server /year, we can not afford, our technology is limited, can not write the cluster solution, so still give up.

Ejabberd Benefits Erlang is born for the cluster multithreading, the cluster is completely free, advanced version is just a few custom development features, is doing more in-depth research work, the shortcomings are also some, less information, we can only rely on the official website, step by step ....

Open Source project Pit is also more, we are slowly familiar with it is a powerful tool, with the bad on the apology of Zigong it haha.

An approach to XMPP server project