Odessa: enabling interactive perception applications on mobile devices

By Moo-Ryong Ra, Anmol Sheth, Lily B. Mummert, Padmanabhan Pillai, David Wetherall, Ramesh Govindan. ACM SIGMOBILE International Conference on Mobile Systems, Applications, and Services, 2011

394 citations recorded by the source corpus.

Subfields: Computer Networks and Communications, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/eb27a6067d7a0416d060434e307e8cf34c11a39f

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Networks and Communications

In Computer Vision and Pattern Recognition

Citations in the corpus

ThinkAir: Dynamic resource allocation and parallel execution in the cloud for mobile code offloading (2012, background): “The basic idea of dynamically switching between (constrained) local and (plentiful) remote resources, often referred as cyber-foraging, has shed light on many research work [3], [4], [5], [6], [7], [8], [9].”
Fast Adaptive Task Offloading in Edge Computing Based on Meta Reinforcement Learning (2020, background): “, face recognition [2], gesture recognition [2], and augmented reality [3]) are composed of dependent tasks, which can bemodelled as a Directed Acyclic Graph (DAG).”
MarkIt: privacy markers for protecting visual secrets (2014, background): “The idea is to offload heavy computation on server/cloud infrastructure [5, 7, 8, 12].”
JointDNN: An Efficient Training and Inference Engine for Intelligent Mobile Cloud Computing Services (2018, background): “There are existing prior arts focusing on offloading computation from the mobile to the cloud[15], [41], [42], [43], [44], [45], [46].”
Computation Offloading Toward Edge Computing (2019, methodology): “Then, graph partitioning algorithms are applied to cut the graph by achieving some goals, such as saving energy [5], [86], [87], reducing execution time [88], minimizing network traffic [89], or maximizing throughput [90].”
DeepDecision: A Mobile Deep Learning Framework for Edge Video Analytics (2018, background): “Mobile offloading: [22] developed a general framework for deciding when to offload, while [23], [24] specifically study interactive visual applications.”
Future Edge Cloud and Edge Computing for Internet of Things Applications (2018, background): “Typical research projects on mobile-edge offloading include MAUI (by Duke University in 2010) [17], CloneCloud (by Intel Labs Berkeley in 2011) [18], Odessa (by University of Southern California in 2011) [19], and COMET (by University of Michigan in 2012) [20].”
Towards wearable cognitive assistance (2014, background): “Odessa [29] makes dynamic offloading and parallelism decisions to achieve low latency for interactive perception ap-”
MCDNN: An Approximation-Based Execution Framework for Deep Stream Processing Under Resource Constraints (2016, background): “Off-loading from mobile device to cloud has long been an option to handle heavyweight computations [16, 19, 38].”
Offloading in Mobile Edge Computing: Task Allocation and Computational Frequency Scaling (2017, methodology): “[24] proposed a dynamic offloading strategy using a greedy algorithm to minimize the completion time of applications.”
Glimpse: Continuous, Real-Time Object Recognition on Mobile Devices (2015, methodology): “Odessa [41] makes offloading and parallelism decisions using runtime profiling with the goal of reducing the total execution delay.”
Neurosurgeon: Collaborative Intelligence Between the Cloud and Mobile Edge (2017, background): “MAUI [34] Comet [35] Odessa [36] CloneCloud [37] Neurosurgeon No need to transfer program state 3 3 Data-centric compute partitioning 3 Low/no runtime overhead 3 3 3 3 Requires no application-specific profiling 3 3 No programmer annotation needed 3 3 3 3 Server load sensitive 3 3”
Optimal Cloudlet Placement and User to Cloudlet Allocation in Wireless Metropolitan Area Networks (2017, background): “Although existing frameworks and algorithms for task offloading mostly focused on clouds [5], [6], cloudlets have been quickly gaining recognition as an alternative offloading destination [7], [8], [9], [10], [11].”
Application-aware traffic scheduling for workload offloading in mobile clouds (2016, background): “Modern mobile devices continuously generate and process sensory data about the surrounding environment, and hence enable users’ contextual awareness [12], [6].”
Privacy-Preserving Inference in Machine Learning Services Using Trusted Execution Environments (2019, background): “Model partitioning and offloading: Previous efforts like [5], [8], [13], [30] focus on improving performance on mobile devices by partitioning and offloading computation.”
A Distributed Truthful Auction Mechanism for Task Allocation in Mobile Cloud Computing (2018, background): “nearby mobile devices through WLAN/WiFi can significantly reduce the communication latency and network congestion [13], [14], [15], [16], [17], [18], [19].”
COMET: Code Offload by Migrating Execution Transparently (2012, neutral): “Other work instead focuses on specific kinds of computation like Odessa [26] which targets perception applications, or SociableSense [27] designed for harvesting social measurements.”
Deep Learning With Edge Computing: A Review (2019, background): “Odessa [114], which divide an application into its constituent subtasks, and decide which subtasks to execute where based on energy and/or latency considerations.”