Microsoft Corporation
Automatic clustering of users for enabling viral adoption of applications hosted by multi-tenant systems

Abstract:

The techniques disclosed herein enable systems to utilize the network effect of end-user viral adoption of collaborative services and applications hosted by multi-tenant computing systems. A system can achieve this by automatically clustering users who independently send requests to generate new tenants. A number of factors can enable a system to cluster users into a single tenant including characteristics of requests to generate new tenants and/or member identities received in association with requests to generate new tenants. Examples of request characteristics can include network addresses indicating a source of each request, a domain name or entity names received in association with a request, a rate in which requests are received, and/or heuristic techniques that utilize various combinations of these and other factors. The automatic selection of users for specific tenants allows a system to allow viral user adoption without a centralized IT manager.