Mobile App Acceleration via Fine-Grain Offloading to the Cloud
| Title: | Mobile App Acceleration via Fine-Grain Offloading to the Cloud |
| Author: |
Lin, Chit-Kwan; Kung, H. T.
Note: Order does not necessarily reflect citation order of authors. |
| Citation: | Lin C-K, and HT Kung. 2014. "Mobile App Acceleration via Fine-Grain Offloading to the Cloud." Presented at the 6th USENIX Workshop on Hot Topics in Cloud Computing (HotCloud '14), June 17-18, 2014, Philadelphia, PA. |
| Full Text & Related Files: |
2014-hotcloud-lin-kung.pdf (316.9Kb; PDF) 
|
| Abstract: | Mobile device hardware can limit the sophistication of mobile applications. One strategy for side-stepping these constraints is to opportunistically offload computations to the cloud, where more capable hardware can do the heavy lifting. We propose a platform that accomplishes this via compressive offloading, a novel application of compressive sensing in a distributed shared memory setting. Our prototype gives up to an order-of-magnitude acceleration and 60% longer battery life to the end user of an example handwriting recognition app. We argue that offloading is beneficial to both end users and cloud providers—the former experiences a performance boost and the latter receives a steady stream of small computations to fill periods of under-utilization. Such workloads, originating from ARM-based mobile devices, are especially well-suited for offloading to emerging ARM-based data centers. |
| Published Version: | https://www.usenix.org/conference/hotcloud14/workshop-program/presentation/lin |
| Terms of Use: | This article is made available under the terms and conditions applicable to Open Access Policy Articles, as set forth at http://nrs.harvard.edu/urn-3:HUL.InstRepos:dash.current.terms-of-use#OAP |
| Citable link to this page: | http://nrs.harvard.edu/urn-3:HUL.InstRepos:12965652 |
| Downloads of this work: |
This item appears in the following Collection(s)
-
FAS Scholarly Articles [15387]
Contact administrator regarding this item (to report mistakes or request changes)
