This method can optimally accomplish data-intensive dispersive computing in a general network, by jointly forwarding computation requests, scheduling computations, and replicating and storing data. The method focuses on data-intensive computation, where computations require significant data inputs. Computation request forwarding decisions, computation scheduling, as well as data request forwarding and caching/storage decisions, are jointly determined.
Technology Overview
Northeastern researchers propose an idea of the DECO (Data-centric Computation) framework for joint computation, caching, and request forwarding in data-centric computing networks. DECO utilizes a virtual control plane which operates on the demand rates for computation and data, and an actual plane which handles computation requests, data requests, data objects and computation results in the physical network. A throughput optimal policy within the virtual plane as a basis for adaptive and distributed computation, caching, and request forwarding in the actual plane.
- Optimizes data-intensive dispersive computation over arbitrary network topologies
- Can jointly determines request forwarding, computation scheduling, and data storage/caching decisions
- Optimality guarantee in terms of the stable throughput of computational requests
- It is both distributed and adaptive
- It can achieve superior performance in terms of request satisfaction delay 
- Mobile edge computing in wireless networks
- Distributed machine learning
- Multiplayer online games
- Distributed databases
- Grid computing
- License
- Partnering
- Research collaboration
Patent Information:
For Information, Contact:
Myron Kassaraba
Director of Commercialization
Northeastern University
Edmund Yeh
Khashayar Kamran
5G Networks
Cellular Networks
Data Storage
Distributed Optimization
Machine Learning
Network Slicing
Next-generation Mobile Networks