Papers by Schahram Dustdar
Computer
This article contributes a research vision for using edge computing to deliver the computing infr... more This article contributes a research vision for using edge computing to deliver the computing infrastructure for emerging smart megacities, with use cases, key requirements, and reflections on the state of the art. We also address edge server placements, a key challenge for edge computing adoption. U rbanization is continuing at unprecedented rates, with estimates suggesting that, by 2030, the global population of major cities with 10 million or more inhabitants will grow from 3.2 billion to almost 5 billion. 1 Parallel to this growth, there is an increasing trend to make cities smart by integrating sensors, networks, and artificial intelligence (AI) to optimize city functions and offer services that support inhabitants' lives. For example, environmental sensors are increasingly being deployed within cities to monitor various pollutants at fine spatial and temporal resolutions,
IEEE/CAA Journal of Automatica Sinica
Cornell University - arXiv, Dec 4, 2021
Multi-server jobs are imperative in modern computing clusters. A multi-server job has multiple ta... more Multi-server jobs are imperative in modern computing clusters. A multi-server job has multiple task components and each of the task components is responsible for processing a specific size of workloads. Efficient online workload dispatching is crucial but challenging to co-located heterogeneous multi-server jobs. The dispatching policy should decide (i) where to launch each task component instance of the arrived jobs and (ii) the size of workloads that each task component processes. Existing policies are explicit and effective when facing service locality and resource contention in both offline and online settings. However, when adding the deadline-aware constraint, the theoretical superiority of these policies could not be guaranteed. To fill the theoretical gap, in this paper, we design an α-competitive online workload dispatching policy for deadline-aware multi-server jobs based on the spatio-temporal resource mesh model. We formulate the problem as a social welfare maximization program and solve it online with several well designed pseudo functions. The social welfare is formulated as the sum of the utilities of jobs and the utility of the computing cluster. The proposed policy is rigorously proved to be α-competitive for some α ≥ 2. We also validate the theoretical superiority of it with simulations and the results show that it distinctly outperforms two handcrafted baseline policies on the social welfare.
IEEE Internet of Things Journal
Deep neural network (DNN) shows great promise in providing more intelligence to ubiquitous end de... more Deep neural network (DNN) shows great promise in providing more intelligence to ubiquitous end devices. However, the existing partition-offloading schemes adopt data-parallel or model-parallel collaboration between devices and the cloud, which does not make full use of the resources of end devices for deep-level parallel execution. This paper proposes eDDNN (i.e. enabling Distributed DNN), a collaborative inference scheme over heterogeneous end devices using cross-platform web technology, moving the computation close to ubiquitous end devices, improving resource utilization, and reducing the computing pressure of data centers. eDDNN implements D2D communication and collaborative inference among heterogeneous end devices with WebRTC protocol, divides the data and corresponding DNN model into pieces simultaneously, and then executes inference almost independently by establishing a layer dependency table. Besides, eDDNN provides a dynamic allocation algorithm based on deep reinforcement learning to minimize latency. We conduct experiments on various datasets and DNNs and further employ eDDNN into a mobile web AR application to illustrate the effectiveness. The results show that eDDNN can achieve the latency decrease by 2.98x, reduce mobile energy by 1.8x, and relieve the computing pressure of the edge server by 2.57x, against a typical partition-offloading approach.
Cornell University - arXiv, Apr 20, 2022
The majority of IoT devices like smartwatches, smart plugs, HVAC controllers, etc., are powered b... more The majority of IoT devices like smartwatches, smart plugs, HVAC controllers, etc., are powered by hardware with a constrained specification (low memory, clock speed and processor) which is insufficient to accommodate and execute large, high-quality models. On such resource-constrained devices, manufacturers still manage to provide attractive functionalities (to boost sales) by following the traditional approach of programming IoT devices/products to collect and transmit data (image, audio, sensor readings, etc.) to their cloud-based ML analytics platforms. For decades, this online approach has been facing issues such as compromised data streams, non-real-time analytics due to latency, bandwidth constraints, costly subscriptions, recent privacy issues raised by users and the GDPR guidelines, etc. In this paper, to enable ultra-fast and accurate AI-based offline analytics on resource-constrained IoT devices, we present an end-toend multi-component model optimization sequence and open-source its implementation. Researchers and developers can use our optimization sequence to optimize high memory, computation demanding models in multiple aspects in order to produce small size, low latency, low-power consuming models that can comfortably fit and execute on resource-constrained hardware. The experimental results show that our optimization components can produce models that are; (i) 12.06 x times compressed; (ii) 0.13% to 0.27% more accurate; (iii) Orders of magnitude faster unit inference at 0.06 ms. Our optimization sequence is generic and can be applied to any state-of-the-art models trained for anomaly detection, predictive maintenance, robotics, voice recognition, and machine vision. INDEX TERMS Edge Intelligence, Neural Networks, Optimization, TinyML, IoT Hardware. I. INTRODUCTION A RTIFICIAL Intelligence (AI) have been used as the principal approach to solve a variety of significant problems in machine translation, video analytics, voice localization, handwriting recognition, etc. Commonly, to provide edge-level AI-functionalities to customers, manufacturers program their IoT devices/products to capture, compress and transmit data (image, audio, sensor readings, etc.) over the network to their central server/cloud where advanced analytics are performed [1]. Although such cloud-based approaches reduce the maintenance cost by keeping the analytics models in one central location, it may not be suitable for most applications [2] because; First, there is a latency caused when transmitting data to a central server for analysis and back to the application. Second, the use of a server for continuous data storage and analysis is expensive because these applications generate high volumes of data. Furthermore, the processing and storage of multiple data streams make the subscription more costly. This design requires a huge amount of reliable bandwidth, which may not always be
2019 IEEE International Conference on Fog Computing (ICFC)
Data Protection is a major research topic concerning the Internet of Things (IoT). IoT systems co... more Data Protection is a major research topic concerning the Internet of Things (IoT). IoT systems continue to permeate deeper into our personal lives, where devices sense, process, and store all kinds of data. This poses various challenges to security and privacy aspects, especially to applications running on resource constrained devices. In this paper we evaluate selected, well established data protection mechanisms that enable confidentiality and integrity of data. Specifically, we look into the performance of different cryptographic block and stream ciphers, hashing algorithms, message authentication codes, signature mechanisms, and key exchange protocols executed on state-ofthe-art resource constrained devices. By providing limitations and data throughput values, our obtained results ease the calculation of performance/data protection thresholds and facilitate the design and development of secure IoT systems.
2017 13th International Conference on Semantics, Knowledge and Grids (SKG)
Complex systems such as Collective Adaptive Systems that include a variety of resources, are incr... more Complex systems such as Collective Adaptive Systems that include a variety of resources, are increasingly being designed to include people in task-execution, and so social computing is not a stand-alone paradigm only, but it is increasingly researched within mixed-resource systems. The Social Computing paradigm has led to significant advancements in engaging people as resources and/services in solving tasks that can not yet be solved by software. Collectives, encapsulating human resources/services, represent one type of an application of social computing, within which people with different type of skills can be engaged to solve one common problem or work on the same project. Mechanisms of managing social collectives are dependent on functional and non-functional parameters of members of social collectives. In this work, we investigate and show experimental results of how provenance data related to those parameters can help better visualize and extract interaction and performance patterns during a collective's run-time. 1 We do not use the term social computing here to refer to social networks but to human-based task execution in complex systems such as CAS.
IEEE Transactions on Sustainable Computing
IEEE INFOCOM 2022 - IEEE Conference on Computer Communications
Employing today's deep neural network (DNN) into the cross-platform web with an offloading way ha... more Employing today's deep neural network (DNN) into the cross-platform web with an offloading way has been a promising means to alleviate the tension between intensive inference and limited computing resources. However, it is still challenging to directly leverage the distributed DNN execution into web apps with the following limitations, including (1) how special computing tasks such as DNN inference can provide fine-grained and efficient offloading in the inefficient JavaScriptbased environment? (2) lacking the ability to balance the latency and mobile energy to partition the inference facing various web applications' requirements. (3) and ignoring that DNN inference is vulnerable to the operating environment and mobile devices' computing capability, especially dedicated web apps. This paper designs AoDNN, an automatic offloading framework to orchestrate the DNN inference across the mobile web and the edge server, with three main contributions. First, we design the DNN offloading based on providing a snapshot mechanism and use multi-threads to monitor dynamic contexts, partition decision, trigger offloading, etc. Second, we provide a learning-based latency and mobile energy prediction framework for supporting various web browsers and platforms. Third, we establish a multiobjective optimization to solve the optimal partition by balancing the latency and mobile energy.
IEEE Transactions on Mobile Computing
Enabling deep learning technology on the mobile web can improve the user's experience for achievi... more Enabling deep learning technology on the mobile web can improve the user's experience for achieving web artificial intelligence in various fields. However, heavy DNN models and limited computing resources of the mobile web are now unable to support executing computationally intensive DNNs when deploying in a cloud computing platform. With the help of promising edge computing, we propose a lightweight collaborative deep neural network for the mobile web, named LcDNN, which contributes to three aspects: (1) We design a composite collaborative DNN that reduces the model size, accelerates inference, and reduces mobile energy cost by executing a lightweight binary neural network (BNN) branch on the mobile web. (2) We provide a jointly training method for LcDNN and implement an energy-efficient inference library for executing the BNN branch on the mobile web. (3) To further promote the resource utilization of the edge cloud, we develop a DRL-based online scheduling scheme to obtain an optimal allocation for LcDNN. The experimental results show that LcDNN outperforms existing approaches for reducing the model size by about 16x to 29x. It also reduces the end-to-end latency and mobile energy cost with acceptable accuracy and improves the throughput and resource utilization of the edge cloud.
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
With the significant improvements in Earth observation (EO) technologies, remote sensing (RS) dat... more With the significant improvements in Earth observation (EO) technologies, remote sensing (RS) data exhibit the typical characteristics of Big Data. Propelled by the powerful feature extraction capabilities of intelligent algorithms, RS image interpretation has drawn remarkable attention and achieved progress. However, the semantic relationship and domain knowledge hidden in massive RS images have not been fully exploited. To the best of our knowledge, a comprehensive review of recent achievements regarding semantic graph-based methods for comprehension and interpretation of RS images is still lacking. Specifically, this article discusses the main challenges of RS image interpretation and presents a systematic survey of typical semantic graph-based methodologies for RS knowledge representation and understanding, including the Ontology Model, Geo-Information Tupu, and Semantic Knowledge Graph. Furthermore, we categorize and summarize how the existing technologies address different challenges in RS image interpretation based on semantic graph-based methods, which indicates that the semantic information about potential relationships and prior knowledge of variant RS targets are central to the solution. In addition, a case study of RS geological interpretation based on the semantic knowledge graph is demonstrated to show the promising capability of intelligent RS image interpretation. Finally, the future directions are discussed for further research.
Blockchain technology offers a sizable promise to rethink the way inter-organizational business p... more Blockchain technology offers a sizable promise to rethink the way inter-organizational business processes are managed because of its potential to realize execution without a central party serving as a single point of trust (and failure). To stimulate research on this promise and the limits thereof, in this paper we outline the challenges and opportunities of blockchain for Business Process Management (BPM). We first reflect how blockchains could be used in the context of the established BPM lifecycle and second how they might become relevant beyond. We conclude our discourse with a summary of seven research directions for investigating the application of blockchain technology in the context of BPM.
This document presents a Services Research Roadmap that launches four pivotal, inherently related... more This document presents a Services Research Roadmap that launches four pivotal, inherently related, research themes to Service-Oriented Computing (SOC): service foundations, service composition, service management and monitoring and service-oriented engineering. Each theme is introduced briefly from a technology, state of the art and scientific challenges standpoint. From the technology standpoint a comprehensive review of state of the art, standards, and current research activities in each key area is provided. From the state of the art the major open problems and bottlenecks to progress are identified. During the during seminar each core theme was initially introduced by a leading expert in the field who described the state of the art and highlighting open problems and important research topics for the SOC community to work on in the future. These experts were then asked to coordinate parallel workgroups that were entrusted with an in-depth analysis of the research opportunities an...
Cornell University - arXiv, Nov 9, 2019
Multi-access Edge Computing (MEC) is booming as a promising paradigm to push the computation and ... more Multi-access Edge Computing (MEC) is booming as a promising paradigm to push the computation and communication resources from cloud to the network edge to provide services and to perform computations. With container technologies, mobile devices with small memory footprint can run composite microservice-based applications without time-consuming backbone. Service placement at the edge is of importance to put MEC from theory into practice. However, current state-of-the-art research does not sufficiently take the composite property of services into consideration. Besides, although Kubernetes has certain abilities to heal container failures, high availability cannot be ensured due to heterogeneity and variability of edge sites. To deal with these problems, we propose a distributed redundant placement framework SAA-RP and a GA-based Server Selection (GASS) algorithm for microservice-based applications with sequential combinatorial structure. We formulate a stochastic optimization problem with the uncertainty of microservice request considered, and then decide for each microservice, how it should be deployed and with how many instances as well as on which edge sites to place them. Benchmark policies are implemented in two scenarios, where redundancy is allowed and not, respectively. Numerical results based on a real-world dataset verify that GASS significantly outperforms all the benchmark policies.
Software's ability to adapt at run-time to changing user needs, system intrusions or faults, ... more Software's ability to adapt at run-time to changing user needs, system intrusions or faults, changing operational environment, and resource variability has been proposed as a means to cope with the complexity of today's software-intensive systems. Such self-adaptive systems can configure and reconfigure themselves, augment their functionality, continually optimize themselves, protect themselves, and recover themselves, while keeping most of their complexity hidden from the user and administrator. In this paper, we present research road map for software engineering of self-adaptive systems focusing on four views, which we identify as essential: requirements, modelling, engineering, and assurances.
Association for Computing Machinery (ACM) Press, Dec 1, 2008
ABSTRACT
IEEE Internet of Things Journal
2009 Eighth International Conference on Grid and Cooperative Computing, Aug 1, 2009
Jiannong Cao, Hong Kong Polytechnic University, China Jinjun Chen, Swinburne University of Techno... more Jiannong Cao, Hong Kong Polytechnic University, China Jinjun Chen, Swinburne University of Technology, Australia Xiaowu Chen, Beihang University, China Xue-bin Chi, Chinese Academy of Sciences, China Qianni Deng, Shanghai Jiaotong University, China Ruihua Di, Beijing University of Technology, China Shoubin Dong, South China University of Technology, China Xiaoshe Dong, Xi'an Jiaotong University, China Xiaoyong Du, Renmin University of China, China Zhihui Du, Tsinghua University, China Schahram Dustdar, Information Systems Institute Vienna ...
Uploads
Papers by Schahram Dustdar