The amount of data processed annually over the Internet has crossed the zetabyte boundary, yet this Big Data
cannot be efficiently processed or stored using today's mobile devices. Parallel to this explosive growth in data, a
substantial increase in mobile compute-capability and the advances in cloud computing have brought the state-of-the-
art in mobile-cloud computing to an inflection point, where the right architecture may allow mobile devices to
run applications utilizing Big Data and intensive computing. In this paper, we propose the MObile Cloud-based
Hybrid Architecture (MOCHA), which formulates a solution to permit mobile-cloud computing applications such
as object recognition in the battlefield by introducing a mid-stage compute- and storage-layer, called the cloudlet.
MOCHA is built on the key observation that many mobile-cloud applications have the following characteristics:
1) they are compute-intensive, requiring the compute-power of a supercomputer, and 2) they use Big Data,
requiring a communications link to cloud-based database sources in near-real-time. In this paper, we describe
the operation of MOCHA in battlefield applications, by formulating the aforementioned mobile and cloudlet to
be housed within a soldier's vest and inside a military vehicle, respectively, and enabling access to the cloud
through high latency satellite links. We provide simulations using the traditional mobile-cloud approach as well
as utilizing MOCHA with a mid-stage cloudlet to quantify the utility of this architecture. We show that the
MOCHA platform for mobile-cloud computing promises a future for critical battlefield applications that access
Big Data, which is currently not possible using existing technology.