Showing 1–7 of 7 results for author: Michalke, M

Evaluating the Impact of Inter-cluster Communications in Edge Computing

Authors: Marc Michalke, Iulisloi Zacarias, Admela Jukan, Kfir Toledo, Etai Lev-Ran

Abstract: Distributed applications based on micro-services in edge computing are becoming increasingly popular due to the rapid evolution of mobile networks. While Kubernetes is the default framework when it comes to orchestrating and managing micro-service-based applications in mobile networks, the requirement to run applications between multiple sites at cloud and edge poses new challenges. Since Kubernet… ▽ More Distributed applications based on micro-services in edge computing are becoming increasingly popular due to the rapid evolution of mobile networks. While Kubernetes is the default framework when it comes to orchestrating and managing micro-service-based applications in mobile networks, the requirement to run applications between multiple sites at cloud and edge poses new challenges. Since Kubernetes does not natively provide tools to abstract inter-cluster communications at the application level, inter-cluster communication in edge computing is becoming increasingly critical to the application performance. In this paper, we evaluate for the first time the impact of inter-cluster communication on edge computing performance by using three prominent, open source inter-cluster communication projects and tools, i.e., Submariner, ClusterLink and Skupper. We develop a fully open-source testbed that integrates these tools in a modular fashion, and experimentally benchmark sample applications, including the ML class of applications, on their performance running in the multi-cluster edge computing system under varying networking conditions. We experimentally analyze two classes of envisioned mobile applications, i.e., a) industrial automation, b) vehicle decision drive assist. Our results show that ClusterLink performs best out of the three tools in scenarios with increased payloads, regardless of the underlying networking conditions or transmission direction between clusters. It is closely followed by Skupper, unless request and reply both transport significant amounts of data. Finally, when requesting smaller amounts of data from a service, Submariner slightly outperforms Skupper and ClusterLink regardless of the inter-node networking conditions. △ Less

Submitted 30 January, 2025; v1 submitted 13 September, 2024; originally announced September 2024.

Comments: Submitted to IEEE/IFIP Network Operations and Management Symposium 2025

Deploying AI-Based Applications with Serverless Computing in 6G Networks: An Experimental Study

Authors: Marc Michalke, Chukwuemeka Muonagor, Admela Jukan

Abstract: Future 6G networks are expected to heavily utilize machine learning capabilities in a wide variety of applications with features and benefits for both, the end user and the provider. While the options for utilizing these technologies are almost endless, from the perspective of network architecture and standardized service, the deployment decisions on where to execute the AI-tasks are critical, esp… ▽ More Future 6G networks are expected to heavily utilize machine learning capabilities in a wide variety of applications with features and benefits for both, the end user and the provider. While the options for utilizing these technologies are almost endless, from the perspective of network architecture and standardized service, the deployment decisions on where to execute the AI-tasks are critical, especially when considering the dynamic and heterogeneous nature of processing and connectivity capability of 6G networks. On the other hand, conceptual and standardization work is still in its infancy, as to how to categorizes ML applications in 6G landscapes; some of them are part of network management functions, some target the inference itself, while many others emphasize model training. It is likely that future mobile services may all be in the AI domain, or combined with AI. This work makes a case for the serverless computing paradigm to be used to this end. We first provide an overview of different machine learning applications that are expected to be relevant in 6G networks. We then create a set of general requirements for software engineering solutions executing these workloads from them and propose and implement a high-level edge-focused architecture to execute such tasks. We then map the ML-serverless paradigm to the case study of 6G architecture and test the resulting performance experimentally for a machine learning application against a setup created in a more traditional, cloud-based manner. Our results show that, while there is a trade-off in predictability of the response times and the accuracy, the achieved median accuracy in a 6G setup remains the same, while the median response time decreases by around 25% compared to the cloud setup. △ Less

Submitted 1 July, 2024; originally announced July 2024.

Comments: Submitted to https://ai-for-6g.com/

An experimental study of the response time in an edge-cloud continuum with ClusterLink

Authors: Marc Michalke, Fin Gentzen, Admela Jukan, Kfir Toledo, Etai Lev Ran

Abstract: In this paper, we conduct an experimental study to provide a general sense of the application response time implications that inter-cluster communication experiences at the edge at the example of a specific IoT-edge-cloud contiuum solution from the EU Project ICOS called ClusterLink. We create an environment to emulate different networking topologies that include multiple cloud or edge sites scena… ▽ More In this paper, we conduct an experimental study to provide a general sense of the application response time implications that inter-cluster communication experiences at the edge at the example of a specific IoT-edge-cloud contiuum solution from the EU Project ICOS called ClusterLink. We create an environment to emulate different networking topologies that include multiple cloud or edge sites scenarios, and conduct a set of tests to compare the application response times via ClusterLink to direct communications in relation to node distances and request/response payload size. Our results show that, in an edge context, ClusterLink does not introduce a significant processing overhead to the communication for small payloads as compared to cloud. For higher payloads and on comparably more aged consumer hardware, ClusterLink version 0.2 introduces communication overhead relative to the delay experienced on the link. △ Less

Submitted 27 May, 2024; originally announced May 2024.

Comments: Submitted to IECCONT workshop co-hosted with Euro-Par 2024 https://2024.euro-par.org/workshops/workshops/

Scaling Performance of Serverless Edge Networking

Authors: Marc Michalke, Francisco Carpio, Admela Jukan

Abstract: When clustering devices at the edge, inter-node latency poses a significant challenge that directly impacts the application performance. In this paper, we experimentally examine the impact that inter-node latency has on application performance by measuring the throughput of an distributed serverless application in a real world testbed. We deploy Knative over a Kubernetes cluster of nodes and emula… ▽ More When clustering devices at the edge, inter-node latency poses a significant challenge that directly impacts the application performance. In this paper, we experimentally examine the impact that inter-node latency has on application performance by measuring the throughput of an distributed serverless application in a real world testbed. We deploy Knative over a Kubernetes cluster of nodes and emulate networking delay between them to compare the performance of applications when deployed over a single-site versus multiple distributed computing sites. The results show that multi-site edge networks achieve half the throughput compared to a deployment hosted at a single site under low processing times conditions, whereas the throughput performance significantly improves otherwise. △ Less

Submitted 27 October, 2023; v1 submitted 13 October, 2023; originally announced October 2023.

Engineering and Experimentally Benchmarking Open Source MQTT Broker Implementations

Authors: Jasenka Dizdarevic, Marc Michalke, Admela Jukan

Abstract: The Message Queuing Telemetry Transport (MQTT) protocol is one of the most widely used IoT protocol solutions. In this work, we are especially interested in open-source MQTT Broker implementations (such as Mosquitto, EMQX, RabbitMQ, VerneMQ, and HiveMQ). To this end, we engineer a network testbed to experimentally benchmark the performance of these implementations in an edge computing context with… ▽ More The Message Queuing Telemetry Transport (MQTT) protocol is one of the most widely used IoT protocol solutions. In this work, we are especially interested in open-source MQTT Broker implementations (such as Mosquitto, EMQX, RabbitMQ, VerneMQ, and HiveMQ). To this end, we engineer a network testbed to experimentally benchmark the performance of these implementations in an edge computing context with constrained devices. In more detail, we engineer an automated deployment and orchestration of the containerized MQTT broker implementations, with support for deployment across either moderately powerful AMD64 devices, or more resource constrained ARM64 devices. The proposed MQTT implementations are evaluated in terms of overhead response time and different payload sizes. Results showed that the hardware platform used as well as the message size, and the network parameters (latency, packet loss and jitter) have a significant impact on the performance differences between the brokers. All results, software tools and code are fully reproducible and free and open source. △ Less

Submitted 23 May, 2023; originally announced May 2023.

Comments: This paper is uploaded here for research community, thus it is for non-commercial purposes

BenchFaaS: Benchmarking Serverless Functions in an Edge Computing Network Testbed

Authors: Francisco Carpio, Marc Michalke, Admela Jukan

Abstract: The serverless computing model has evolved as one of the key solutions in the cloud for fast autoscaling and capacity planning. In edge computing environments, however, the serverless model is challenged by the system heterogeneity and performance variability. In this paper, we introduce BenchFaaS, an open-source edge computing network testbed which automates the deployment and benchmarking of ser… ▽ More The serverless computing model has evolved as one of the key solutions in the cloud for fast autoscaling and capacity planning. In edge computing environments, however, the serverless model is challenged by the system heterogeneity and performance variability. In this paper, we introduce BenchFaaS, an open-source edge computing network testbed which automates the deployment and benchmarking of serverless functions. Our edge computing network considers a cluster of virtual machines and Raspberry Pis, and is designed to benchmark serverless functions under different hardware and network conditions. We measure and evaluate: (i) overhead incurred by testbed, (ii) performance of compute intensive tasks, (iii) impact of application payload size, (iv) scalability, and (v) performance of chained serverless functions. We share the lessons learnt in engineering and implementing the testbed. We present the measurement results and analyze the impact of networked infrastructure on serverless performance. The measurements indicate that a properly dimensioned edge computing network can effectively serve as a serverless infrastructure. △ Less

Submitted 17 August, 2022; v1 submitted 5 June, 2022; originally announced June 2022.

Engineering and Experimentally Benchmarking a Serverless Edge Computing System

Authors: Francisco Carpio, Marc Michalke, Admela Jukan

Abstract: Thanks to the latest advances in containerization, the serverless edge computing model is becoming close to reality. Serverless at the edge is expected to enable low latency applications with fast autoscaling mechanisms, all running on heterogeneous and resource-constrained devices. In this work, we engineer and experimentally benchmark a serverless edge computing system architecture. We deploy a… ▽ More Thanks to the latest advances in containerization, the serverless edge computing model is becoming close to reality. Serverless at the edge is expected to enable low latency applications with fast autoscaling mechanisms, all running on heterogeneous and resource-constrained devices. In this work, we engineer and experimentally benchmark a serverless edge computing system architecture. We deploy a decentralized edge computing platform for serverless applications providing processing, storage, and communication capabilities using only open-source software, running over heterogeneous resources (e.g., virtual machines, Raspberry Pis, or bare metal servers, etc). To achieve that, we provision an overlay-network based on Nebula network agnostic technology, running over private or public networks, and use K3s to provide hardware abstraction. We benchmark the system in terms of response times, throughput and scalability using different hardware devices connected through the public Internet. The results show that while serverless is feasible on heterogeneous devices showing a good performance on constrained devices, such as Raspberry Pis, the lack of support when determining computational power and network characterization leaves much room for improvement in edge environments. △ Less

Submitted 11 May, 2021; originally announced May 2021.