ACM/IFIP/USENIX Middleware 2017

Las Vegas, Nevada

Dec 11-15, 2017

Doctoral Symposium

The following papers will be presented at the pre-conference on Monday (11th Dec). Click on each title below to see the abstract. The ACM link to the papers have also been added and the PDFs will be available for one month starting from the first day of the conference (Dec 11th).

Sharing a connectivity through informed proxy selection process (link)

Khulan Batbayar (Universitat Politècnica de Catalunya)

Approximately half of the world is living without a reliable In- ternet connection. Many rural, distant areas are managing their network connectivity through sharing an available Internet con- nection within the community with the use of proxy server nodes. Proxy nodes have limited capacity and in Wireless Mesh Network (WMN) they often overloaded with client requests. In order to mini- mize and distribute the load over the proxy nodes, we are suggesting an informed proxy selection process for the client nodes with the help of randomized fair selection mechanism.

Message-Oriented Middleware for Edge Computing Applications (link)

Thomas Rausch (TU Wien)

Edge computing is an emerging paradigm which aims to leverage the ever increasing amount of computational resources at the edge of the network to satisfy the stringent quality of service (QoS) re- quirements of many modern Internet of Things (IoT) scenarios. This PhD thesis explores challenges and solutions of message-oriented middleware (MOM) for edge computing applications. In particular, we focus on QoS optimization and message delivery guarantees under the constraints of geographic dispersion, client mobility, dynamic resource availability, and privacy policies.

Continuous Experimentation for Software Developers (link)

Gerald Schermann (University of Zurich)

Recent advances in build, test, and deployment automation not only enable companies shipping new functionality faster to their users, but also provide them the ability to experiment with functionality on small fractions of the user base rst. These experiments involve techniques such as A/B testing, canary releases, or dark launches. However, neither managing multiple experiments in parallel (i.e., operating and monitoring multiple versions), nor specifying param- eters for experiments (e.g., to avoid that they negatively impact each other) is a trivial task. In my research, I want to support developers and release engineers conducting experiments in an automated and data-driven way.

End-to-End Regression Testing for Distributed Systems (link)

Eugenia Gabrielova (University of California, Irvine)

Even with substantial advances in tools and research techniques, distributed systems remain challenging to test. One frustrating aspect of distributed systems development is the resurfacing of old problems due to code changes. Regression test suites replicate previously known bugs and ensure they do not resurface as the code evolves. Conventional unit regression tests miss a substantial amount of distributed system problems; end-to-end testing is almost always required in order to reproduce complex bugs. We describe a framework for regression testing that bridges a gap between local ad-hoc experiments and end-to-end stress testing, potentially lowering the recurrence of critical bugs.

Towards a Framework for Orchestrated Distributed Database Evaluation in the Cloud (link)

Daniel Seybold (Ulm University)

The selection and operation of a distributed database management system (DDBMS) in the cloud is a challenging task as supportive evaluation frameworks miss orchestrated evaluation scenarios, hindering comparable and reproducible evaluations for heterogeneous cloud resources. We propose a novel evaluation approach that supports orchestrated evaluation scenarios for scalability, elasticity and availability by exploiting cloud resources. We highlight the challenges in evaluating DDBMSs in the cloud and introduce a cloud-centric framework for orchestrated DDBMS evaluation, enabling reproducible evaluations and significant rating indices.

Toward Software Updates in IoT Environments: Why Existing P2P Systems are not Enough (link)

Trystram (INSA Lyon - Red Hat)

The number of connected devices is growing, as well as their embedded software complexity. Applications require man- agement and updates. Modern software orchestrators and management systems are mostly centralised and expensive to scale to very large systems. We propose a decentralised approach to distribute software updates more e ciently to IoT devices using P2P. We seek to adapt system behaviour to IoT constraints, such as device heterogeneity, unreliable network connectivity and applications speci cs.

Towards Accelerating Synchrophasor Based Linear State Estimation of Power Grid Systems (link)

Vinaya Chakati (Arizona State University)

Phasor Measurement Units (PMUs) are high speed monitoring de- vices that present a reliable and dynamic picture of the power grid. Many real time grid applications may bene t from these measure- ments. Synchrophasor only Linear State Estimator (LSE) is one of the critical Energy Management System (EMS) application that is bene ted from the PMU measurements. Increase in the number of PMUs and grid size; increases computational burden of the LSE solver. Installing additional hardware may be a possible solution to deal with computational burden. However this incurs huge in- frastructure, operation and maintenance cost. This paper, presents a cost e ective cloud computing solution to address the compu- tational burden of the LSE solver. Further, we plan to extend this work to address the limitations encountered by the Cloud hosted LSE solver.