A cloudlet is a small-scale data center or cluster of computers designed to quickly provide cloud computing services to mobile devices, such as smartphones, tablets and wearable devices, within close geographical proximity.
The goal of a cloudlet is to increase the response time of applications running on mobile devices by using low latency, high-bandwidth wireless connectivity and by hosting cloud computing resources, such as virtual machines, physically closer to the mobile devices accessing them. This is intended to eliminate the wide area network (WAN) latency delays that can occur in traditional cloud computing models.
The cloudlet was specifically designed to support interactive and resource-intensive mobile applications, such as those for speech recognition, language processing, machine learning and virtual reality.
Key differences between a cloudlet and a public cloud data center
A cloudlet is considered a form of cloud computing because it delivers hosted services to users over a network. However, a cloudlet differs from a public cloud data center, such as those operated by public cloud providers like Amazon Web Services, in a number of ways.
First, a cloudlet is self-managed by the businesses or users that employ it, while a public cloud data center is managed full-time by a cloud provider. Second, a cloudlet predominantly uses a local area network (LAN) for connectivity, versus the public Internet. Thirdly, a cloudlet is employed by fewer, more localized users than a major public cloud service. Finally, a cloudlet contains only "soft state" copies of data, such as a cache copy, or code that is stored elsewhere.
The cloudlet prototype
A prototype implementation of a cloudlet was originally developed by Carnegie Mellon University as a research project, starting in 2009. The term cloudlet was coined by computer scientists Mahadev Satyanarayanan, Victor Bahl, Ramón Cáceres and Nigel Davies.