Umakishore Ramachandran
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Umakishore Ramachandran has a total of 68 co-authors. Here are the top co-authors by number of shared publications:
- Nova Ahmed
Design Inclusion and Access Lab (DIAL), North South University, Dhaka, Bangladesh
3 shared publications - Summit Shrestha
Georgia Institute of Technology, Atlanta, Georgia, USA
3 shared publications - Alex Cabral
Massachusetts Institute of Technology, Cambridge, Massachusetts, USA
2 shared publications - Anirudh Sarma
College of Computing, Georgia Tech, Atlanta, Georgia, USA
2 shared publications - Ashutosh Dhekne
Georgia Tech, Atlanta, Georgia, USA
2 shared publications - Bikash Agarwalla
Georgia Institute of Technology
2 shared publications
Latest Publications
- Open Access
- 2 September 2026
- Published by Association for Computing Machinery (ACM)
- p. 19-27
LM4D: Exploring Telephony-Mediated Access to Language Models for Underserved Communities
Language Models (LMs) are increasingly used for information access, contextual question answering, and agentic workflows. However, most LM interfaces require modern smartphones or browsers, reliable internet connectivity, and paid cloud subscriptions, limiting access for users in low-resource and underserved communities. This paper explores how LM access can be delivered using infrastructure already widely available in underserved regions with minimal to no cognitive load. We present LM4D (Language Models for Developing Communities), a voice-first plug-and-play architecture that enables users to interact with LMs through ordinary phone calls without requiring Internet connectivity or modern devices. The system integrates telephony streaming, speech-to-text, and text-to-speech pipelines with an optional language-aware routing mechanism, a lightweight governance layer with identity management and usage quotas, and flexible backend connectivity to either cloud-hosted or locally deployed language models. We implement a prototype for an educational use case and evaluate its feasibility under resource-constrained underserved contexts by analyzing latency, response quality, and operational cost. Our findings show early architectural feasibility of a telephony-mediated approach to language model access for underserved communities. Yet sociotechnical dimensions such as usability, affordability, community governance, and low-resource language support remain open challenges in helping reduce the digital divide in AI access, motivating future work.
- 2 September 2026
- Published by Association for Computing Machinery (ACM)
- p. 19-27
Open AccessLM4D: Exploring Telephony-Mediated Access to Language Models for Underserved Communities
Language Models (LMs) are increasingly used for information access, contextual question answering, and agentic workflows. However, most LM interfaces require modern smartphones or browsers, reliable internet connectivity, and paid cloud subscriptions, limiting access for users in low-resource and underserved communities. This paper explores how LM access can be delivered using infrastructure already widely available in underserved regions with minimal to no cognitive load. We present LM4D (Language Models for Developing Communities), a voice-first plug-and-play architecture that enables users to interact with LMs through ordinary phone calls without requiring Internet connectivity or modern devices. The system integrates telephony streaming, speech-to-text, and text-to-speech pipelines with an optional language-aware routing mechanism, a lightweight governance layer with identity management and usage quotas, and flexible backend connectivity to either cloud-hosted or locally deployed language models. We implement a prototype for an educational use case and evaluate its feasibility under resource-constrained underserved contexts by analyzing latency, response quality, and operational cost. Our findings show early architectural feasibility of a telephony-mediated approach to language model access for underserved communities. Yet sociotechnical dimensions such as usability, affordability, community governance, and low-resource language support remain open challenges in helping reduce the digital divide in AI access, motivating future work.
- Open Access
- 14 December 2025
- Published by Association for Computing Machinery (ACM)
- p. 469-481
A Hybrid Runtime for Function-as-a-Service at the Edge
- 14 December 2025
- Published by Association for Computing Machinery (ACM)
- p. 469-481
Open AccessA Hybrid Runtime for Function-as-a-Service at the Edge
- Open Access
- 3 December 2025
- Published by Association for Computing Machinery (ACM)
- p. 1-8
DroEdgeEM: A Drone-Edge Collaborative Emulation Platform for Emerging Situation-aware Applications
- 3 December 2025
- Published by Association for Computing Machinery (ACM)
- p. 1-8
Open AccessDroEdgeEM: A Drone-Edge Collaborative Emulation Platform for Emerging Situation-aware Applications
- Open Access
- 21 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 647-665
Bringing Context to the Underserved: Rethinking Context-Aware Design to Bridge the Digital Divide
- 21 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 647-665
Open AccessBringing Context to the Underserved: Rethinking Context-Aware Design to Bridge the Digital Divide
- Open Access
- 24 June 2024
- Published by Association for Computing Machinery (ACM)
- p. 78-89
FEO: Efficient Resource Allocation for FaaS at the Edge
Geo-distributed Edge sites are expected to cater to the stringent demands of situation-aware applications like collaborative autonomous vehicles and drone swarms. While clients of such applications benefit from having network-proximal compute resources, an Edge site has limited resources compared to the traditional Cloud. Moreover, the load experienced by an Edge site depends on a client's mobility pattern, which may often be unpredictable. The Function-as-a-Service (FaaS) paradigm is poised aptly to handle the ephemeral nature of workload demand at Edge sites. In FaaS, applications are decomposed into containerized functions enabling fine-grained resource management. However, spatio-temporal variations in client mobility can still lead to rapid saturation of resources beyond the capacity of an Edge site. To address this challenge, we develop FEO (Federated Edge Orchestrator), a resource allocation scheme across the geodistributed Edge infrastructure for FaaS. FEO employs a novel federated policy to offload function invocations to peer sites with spare resource capacity without the need to frequently share knowledge about available capacities among participating sites. Detailed experiments show that FEO's approach can reduce a site's P99 latency by almost 3x, while maintaining application service level objectives at all other sites.
- 24 June 2024
- Published by Association for Computing Machinery (ACM)
- p. 78-89
Open AccessFEO: Efficient Resource Allocation for FaaS at the Edge
Geo-distributed Edge sites are expected to cater to the stringent demands of situation-aware applications like collaborative autonomous vehicles and drone swarms. While clients of such applications benefit from having network-proximal compute resources, an Edge site has limited resources compared to the traditional Cloud. Moreover, the load experienced by an Edge site depends on a client's mobility pattern, which may often be unpredictable. The Function-as-a-Service (FaaS) paradigm is poised aptly to handle the ephemeral nature of workload demand at Edge sites. In FaaS, applications are decomposed into containerized functions enabling fine-grained resource management. However, spatio-temporal variations in client mobility can still lead to rapid saturation of resources beyond the capacity of an Edge site. To address this challenge, we develop FEO (Federated Edge Orchestrator), a resource allocation scheme across the geodistributed Edge infrastructure for FaaS. FEO employs a novel federated policy to offload function invocations to peer sites with spare resource capacity without the need to frequently share knowledge about available capacities among participating sites. Detailed experiments show that FEO's approach can reduce a site's P99 latency by almost 3x, while maintaining application service level objectives at all other sites.
- Open Access
- 24 June 2024
- Published by Association for Computing Machinery (ACM)
- p. 123-134
Color-based Lightweight Utility-aware Load Shedding for Real-Time Video Analytics at the Edge
Real-time video analytics typically require video frames to be processed by a query to identify objects or activities of interest while adhering to an end-to-end frame processing latency constraint. This imposes a continuous and heavy load on backend compute and network infrastructure. Video data has inherent redundancy and does not always contain an object of interest for a given query. We leverage this property of video streams to propose a lightweight Load Shedder that can be deployed on edge servers or on inexpensive edge devices co-located with cameras. The proposed Load Shedder uses pixel-level color-based features to calculate a utility score for each ingress video frame and a minimum utility threshold to select interesting frames to send for query processing. Dropping unnecessary frames enables the video analytics query in the backend to meet the end-to-end latency constraint with fewer compute and network resources. To guarantee a bounded end-to-end latency at runtime, we introduce a control loop that monitors the backend load and dynamically adjusts the utility threshold. Performance evaluations show that the proposed Load Shedder selects a large portion of frames containing each object of interest while meeting the end-to-end frame processing latency constraint. Furthermore, it does not impose a significant latency overhead when running on edge devices with modest compute resources.
- 24 June 2024
- Published by Association for Computing Machinery (ACM)
- p. 123-134
Open AccessColor-based Lightweight Utility-aware Load Shedding for Real-Time Video Analytics at the Edge
Real-time video analytics typically require video frames to be processed by a query to identify objects or activities of interest while adhering to an end-to-end frame processing latency constraint. This imposes a continuous and heavy load on backend compute and network infrastructure. Video data has inherent redundancy and does not always contain an object of interest for a given query. We leverage this property of video streams to propose a lightweight Load Shedder that can be deployed on edge servers or on inexpensive edge devices co-located with cameras. The proposed Load Shedder uses pixel-level color-based features to calculate a utility score for each ingress video frame and a minimum utility threshold to select interesting frames to send for query processing. Dropping unnecessary frames enables the video analytics query in the backend to meet the end-to-end latency constraint with fewer compute and network resources. To guarantee a bounded end-to-end latency at runtime, we introduce a control loop that monitors the backend load and dynamically adjusts the utility threshold. Performance evaluations show that the proposed Load Shedder selects a large portion of frames containing each object of interest while meeting the end-to-end frame processing latency constraint. Furthermore, it does not impose a significant latency overhead when running on edge devices with modest compute resources.
- Open Access
- 18 June 2023
- Published by Association for Computing Machinery (ACM)
- p. 1-5
EVA: An End-to-End Exploratory Video Analytics System
In recent years, deep learning models have revolutionized computer vision, enabling diverse applications. However, these models are computationally expensive, and leveraging them for video analytics involves low-level imperative programming. To address these efficiency and usability challenges, the database community has developed video database management systems (VDBMSs). However, existing VDBMSs lack extensibility and composability and do not support holistic system optimizations, limiting their practical application. In response to these issues, we present our vision for EVA, a VDBMS that allows for extensible support of user-defined functions and employs a Cascades-style query optimizer. Additionally, we leverage Ray's distributed execution to enhance scalability and performance and explore hardware-specific optimizations to facilitate runtime optimizations. We discuss the architecture and design of EVA, our achievements thus far, and our research roadmap.
- 18 June 2023
- Published by Association for Computing Machinery (ACM)
- p. 1-5
Open AccessEVA: An End-to-End Exploratory Video Analytics System
In recent years, deep learning models have revolutionized computer vision, enabling diverse applications. However, these models are computationally expensive, and leveraging them for video analytics involves low-level imperative programming. To address these efficiency and usability challenges, the database community has developed video database management systems (VDBMSs). However, existing VDBMSs lack extensibility and composability and do not support holistic system optimizations, limiting their practical application. In response to these issues, we present our vision for EVA, a VDBMS that allows for extensible support of user-defined functions and employs a Cascades-style query optimizer. Additionally, we leverage Ray's distributed execution to enhance scalability and performance and explore hardware-specific optimizations to facilitate runtime optimizations. We discuss the architecture and design of EVA, our achievements thus far, and our research roadmap.
- 10 May 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in Computer
- Vol. 54 (5) , 24-33
eCloud: A Vision for the Evolution of the Edge-Cloud Continuum
We present a holistic vision for the edge-cloud ecosystem, with the intent of spurring the creation of next-generation technologies for futuristic applications that operate at computational-perception speeds to convert sensed data to actionable knowledge.
- 10 May 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in Computer
- Vol. 54 (5) , 24-33
eCloud: A Vision for the Evolution of the Edge-Cloud Continuum
We present a holistic vision for the edge-cloud ecosystem, with the intent of spurring the creation of next-generation technologies for futuristic applications that operate at computational-perception speeds to convert sensed data to actionable knowledge.
- Open Access
- 11 June 2012
- Published by Association for Computing Machinery (ACM)
- Vol. 40 (1) , 235-246
What is a good buffer cache replacement scheme for mobile flash storage?
- 11 June 2012
- Published by Association for Computing Machinery (ACM)
- Vol. 40 (1) , 235-246
Open AccessWhat is a good buffer cache replacement scheme for mobile flash storage?