2022 IEEE International Conference on Autonomic Computing and Self-Organizing Systems (ACSOS)最新文献

{"title":"Towards a Framework for Adapting Machine Learning Components","authors":"Maria Casimiro, P. Romano, D. Garlan, Luís Rodrigues","doi":"10.1109/ACSOS55765.2022.00031","DOIUrl":"https://doi.org/10.1109/ACSOS55765.2022.00031","url":null,"abstract":"Machine Learning (ML) models are now commonly used as components in systems. As any other component, ML components can produce erroneous outputs that may penalize system utility. In this context, self-adaptive systems emerge as a natural approach to cope with ML mispredictions, through the execution of adaptation tactics such as model retraining. To synthesize an adaptation strategy, the self-adaptation manager needs to reason about the cost-benefit tradeoffs of the applicable tactics, which is a non-trivial task for tactics such as model retraining, whose benefits are both context- and data-dependent.To address this challenge, this paper proposes a probabilistic modeling framework that supports automated reasoning about the cost/benefit tradeoffs associated with improving ML components of ML-based systems. The key idea of the proposed approach is to decouple the problems of (i) estimating the expected performance improvement after retrain and (ii) estimating the impact of ML improved predictions on overall system utility.We demonstrate the application of the proposed framework by using it to self-adapt a state-of-the-art ML-based fraud-detection system, which we evaluate using a publicly-available, real fraud detection dataset. We show that by predicting system utility stemming from retraining a ML component, the probabilistic model checker can generate adaptation strategies that are significantly closer to the optimal, as compared against baselines such as periodic retraining, or reactive retraining.","PeriodicalId":308147,"journal":{"name":"2022 IEEE International Conference on Autonomic Computing and Self-Organizing Systems (ACSOS)","volume":"100 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124426585","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Message from the Special Event Chairs","authors":"","doi":"10.1109/acsos55765.2022.00008","DOIUrl":"https://doi.org/10.1109/acsos55765.2022.00008","url":null,"abstract":"","PeriodicalId":308147,"journal":{"name":"2022 IEEE International Conference on Autonomic Computing and Self-Organizing Systems (ACSOS)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121543162","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Message from the Program Chairs","authors":"SmartCloud","doi":"10.1109/SCC.2005.64","DOIUrl":"https://doi.org/10.1109/SCC.2005.64","url":null,"abstract":"I extend my most sincere welcome to all attendees of this year's 2016 The IEEE International Conference on Smart Cloud (SmartCloud 2016) hosted at Columbia University, New York, USA. SmartCloud 2016 has involved a variety of topics focusing smart cloud fields presented by authors from distinct research or higher education institutions. The contributions and concentrations have brought a plurality of interests and insights to our picturesque meeting venue. I hope that every attendees and contributors of this conference can take advantage of this opportunity through all activities held by this event, from presentations to keynote lectures. Moreover, it is our great honor to introduce the program for the SmartCloud 2016. We received over 190 technical papers for this conference out of which fewer than 56 were accepted as full papers. We have put together a few tracks over the two-day period of the conference. As a conference chair of SmartCloud 2016, I know that the success of the conference depends ultimately on many people who have endeavored to the conference with us in planning and organizing both the technical programs and social arrangements. We particularly express our gratitude to: General Chairs Sun-Yuan Kung, Qing Yang, and Meikang Qiu; Program Chairs Yongxin Zhu, Yan Zhang, Lixin Tao; Operation Chair Keke Gai, Publicity Chairs Bo Yuan; Local Chair Xiaofu He; Finance Chair Wenyun Dai. We also owe special thanks to our sponsors: IEEE, IEEE Computer Society, IEEE TCSC, Pace University, Columbia University, Stony Brook University, Shanghai Jiao Tong University, Longxiang High-Tech, IEEE Transactions on Computers, IEEE Transactions on Cloud Computing, and North America Chinese Talents Association. This year SmartCloud received 194 submissions. We highly appreciate the effort of all the authors in submitting their papers to SmartCloud 2016. All submissions were reviewed thoroughly by the program committee. Each submission received at least three reviews. After the review process, 48 papers were accepted as full papers. The acceptance rate is 24.74% On behalf of the SmartCloud 2016 organizing committee, I am honored and delighted to welcome you to The IEEE International Conference on Smart Cloud at New York, USA. I believe we have chosen a venue in which guarantees a successful technical conference amid the culture and scenery of Metropolitan New York City. With our charming venue, we are sure that you will enjoy your journey at SmartCloud 2016 for socializing with other scholars and gridding the academic social networks. We also hope you enjoy your stay in New York.","PeriodicalId":308147,"journal":{"name":"2022 IEEE International Conference on Autonomic Computing and Self-Organizing Systems (ACSOS)","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121767272","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Towards Antifragility in Contested Environments: Using Adversarial Search to Learn, Predict, and Counter Open-Ended Threats","authors":"S. S. Hashmi, K. Dam, P. Smet, Mohan Baruwal Chhetri","doi":"10.1109/ACSOS55765.2022.00032","DOIUrl":"https://doi.org/10.1109/ACSOS55765.2022.00032","url":null,"abstract":"Resilience and antifragility under duress present significant challenges for autonomic and self-adaptive systems operating in contested environments. In such settings, the system has to continually plan ahead, accounting for either an adversary or an environment that may negate its actions or degrade its capabilities. This will involve projecting future states, as well as assessing recovery options, counter-measures, and progress towards system goals. For antifragile systems to be effective, we envision three self-* properties to be of key importance: self-exploration, self-learning and self-training. Systems should be able to efficiently self-explore – using adversarial search – the potential impact of the adversary’s attacks and compute the most resilient responses. The exploration can be assisted by prior knowledge of the adversary’s capabilities and attack strategies, which can be self-learned – using opponent modelling – from previous attacks and interactions. The system can self-train – using reinforcement learning – such that it evolves and improves itself as a result of being attacked. This paper discusses those visions and outlines their realisation in AWaRE, a cyber-resilient and self-adaptive multi-agent system.","PeriodicalId":308147,"journal":{"name":"2022 IEEE International Conference on Autonomic Computing and Self-Organizing Systems (ACSOS)","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127974545","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"SHIL: Self-Supervised Hybrid Learning for Security Attack Detection in Containerized Applications","authors":"Yuhang Lin, Olufogorehan Tunde-Onadele, Xiaohui Gu, Jingzhu He, Hugo Latapie","doi":"10.1109/ACSOS55765.2022.00022","DOIUrl":"https://doi.org/10.1109/ACSOS55765.2022.00022","url":null,"abstract":"Container security has received much research attention recently. Previous work has proposed to apply various machine learning techniques to detect security attacks in containerized applications. On one hand, supervised machine learning schemes require sufficient labelled training data to achieve good attack detection accuracy. On the other hand, unsupervised machine learning methods are more practical by avoiding training data labelling requirements, but they often suffer from high false alarm rates. In this paper, we present SHIL, a self-supervised hybrid learning solution, which combines unsupervised and supervised learning methods to achieve high accuracy without requiring any manual data labelling. We have implemented a prototype of SHIL and conducted experiments over 41 real world security attacks in 28 commonly used server applications. Our experimental results show that SHIL can reduce false alarms by 39-91% compared to existing supervised or unsupervised machine learning schemes while achieving a higher or similar detection rate.","PeriodicalId":308147,"journal":{"name":"2022 IEEE International Conference on Autonomic Computing and Self-Organizing Systems (ACSOS)","volume":"91 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121160758","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Self-stabilising Priority-Based Multi-Leader Election and Network Partitioning","authors":"Danilo Pianini, Roberto Casadei, Mirko Viroli","doi":"10.1109/ACSOS55765.2022.00026","DOIUrl":"https://doi.org/10.1109/ACSOS55765.2022.00026","url":null,"abstract":"A common task in situated distributed systems is the self-organising election of leaders. These leaders can be devices or software agents appointed, for instance, to coordinate the activities of other agents or processes. In this work, we focus on the multi-leader election problem in networks of asynchronous message-passing devices, which are a common model in self-organisation approaches like aggregate computing. Specifically, we introduce a novel algorithm for space- and priority-based leader election and compare it with the state of the art. We call the algorithm Bounded Election since it leverages bounding (i.e. minimisation or maximisation) of candidacy messages to drop or promote candidate leaders and ensure stabilisation. The proposed algorithm is formally proven to be self-stabilising, allows for leader prioritisation, and performs on-the-fly network partitioning (namely, as a side effect of the leader election process, the areas regulated by the leaders are also established). Also, we experimentally compare its performance together with the state of the art of leader election in aggregate computing in a variety of synthetic scenarios, showing benefits in terms of convergence time and resilience.","PeriodicalId":308147,"journal":{"name":"2022 IEEE International Conference on Autonomic Computing and Self-Organizing Systems (ACSOS)","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121530574","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Message from the Posters and Demos Chairs","authors":"","doi":"10.1109/acsos55765.2022.00009","DOIUrl":"https://doi.org/10.1109/acsos55765.2022.00009","url":null,"abstract":"","PeriodicalId":308147,"journal":{"name":"2022 IEEE International Conference on Autonomic Computing and Self-Organizing Systems (ACSOS)","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125772843","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"BLOC: Balancing Load with Overload Control In the Microservices Architecture","authors":"Ratnadeep Bhattacharya, Timothy Wood","doi":"10.1109/ACSOS55765.2022.00027","DOIUrl":"https://doi.org/10.1109/ACSOS55765.2022.00027","url":null,"abstract":"The microservices architecture has become ubiquitous in the cloud environment. It simplifies application development by breaking monolithic applications into manageable micro services that can be developed and deployed independently of the whole. However, the move from a monolithic or simple multi-tier architecture to a distributed microservice \"service mesh\" leads to new challenges due to the more complex application topology.A particular problem when automatically managing the performance of microservices is that since each service component scales up and down independently, it can easily create load imbalance problems on shared backend services accessed by multiple components. Traditional load balancing algorithms were designed for centralized load balancers sitting between a group of clients and a server farm. These algorithms, however, do not port over well to a distributed microservice architecture where load balancers are deployed client-side. In this paper we propose a self managing load balancing system, BLOC, which provides consistent response times to users without using a centralized metadata store or explicit messaging between nodes.We show that different service layers scaling independently can create unacceptably wide response time distributions and long tails, hurting client experience. This is because popular microservice load balancing algorithms, like Least Connection, only use a single component’s view of the backend load to guide decisions. This limited perspective leads to an unevenly balanced system and the potential for incast problems where a large number of frontend components can easily overload a shared backend. BLOC uses overload control approaches like rate limiting, active queue management and backpressure to provide feedback to the load balancers. The load balancers react to this feedback with techniques like backoff and retries. We show that this performs significantly better in solving the incast problem in microservice architectures.Evaluating this framework, we found that BLOC improves the response time distribution range, between the 10th and 90th percentiles, by 2 to 4 times and the tail, 99th percentile, latency by two times.","PeriodicalId":308147,"journal":{"name":"2022 IEEE International Conference on Autonomic Computing and Self-Organizing Systems (ACSOS)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129855190","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Reducing the Tail Latency of Microservices Applications via Optimal Configuration Tuning","authors":"Gagan Somashekar, Amoghavarsha Suresh, Saurabh Tyagi, Vikas Dhyani, Krishna Chaitanya Donkada, Anurag Pradhan, Anshul Gandhi","doi":"10.1109/ACSOS55765.2022.00029","DOIUrl":"https://doi.org/10.1109/ACSOS55765.2022.00029","url":null,"abstract":"The microservice architecture is an architectural style for designing applications that supports a collection of fine-grained and loosely-coupled services, called microservices, enabling independent development and deployment. An undesirable complexity that results from this style is the large state space of possibly inter-dependent configuration parameters (of the constituent microservices) which have to be tuned to improve application performance.This paper investigates optimization algorithms to address the problem of configuration tuning of microservices applications. To address the critical issue of large state space, practical dimensionality reduction strategies are developed based on available system characteristics. The evaluation of the optimization algorithms and dimensionality reduction techniques across three popular benchmarking applications highlights the importance of configuration tuning to reduce tail latency (by as much as 46%). A detailed analysis of the efficacy of different dimensionality reduction techniques in capturing the most important parameters is performed using ANOVA techniques. Results show that the right combination of optimization algorithms and dimensionality reduction can provide substantial latency improvements by identifying the right subset of parameters to tune, reducing the search space by as much as 83%.","PeriodicalId":308147,"journal":{"name":"2022 IEEE International Conference on Autonomic Computing and Self-Organizing Systems (ACSOS)","volume":"310 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134473631","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Systematic Review of Fault Tolerance Techniques for Adaptive and Context-Aware Systems","authors":"Kathiani E. Souza, F. Ferrari","doi":"10.1109/ACSOS55765.2022.00020","DOIUrl":"https://doi.org/10.1109/ACSOS55765.2022.00020","url":null,"abstract":"Adaptive systems (ASs) and context-aware systems (CASs) are able to automatically modify themselves in response to changes in their environment. Fault tolerance techniques are used to allow a system to deliver a correct service even in the presence of faults. With respect to ASs and CASs, all environmental conditions may never be tested during the development phase, which can increase the risk of failures. Thus, it is important to apply fault tolerance techniques to recover these systems from error states and continue the delivery of service. This paper reports on results of a systematic literature review that characterises the state-of-the-art of fault tolerance techniques that have been proposed for, or applied to, ASs and CASs. Based on 41 selected studies we: grouped the studies among 7 types of techniques; identified 14 faults; and classified these faults in 4 generic fault types. We found different types of fault tolerance techniques, such as error recovery and diversity. Moreover, we identified faults, such as faults in sensors and actuators. Our results may provide a guidance for developers and researchers for the application of fault tolerance techniques in ASs and CASs.","PeriodicalId":308147,"journal":{"name":"2022 IEEE International Conference on Autonomic Computing and Self-Organizing Systems (ACSOS)","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129669832","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}