Containers are lightweight OS-level virtualization. In recent years, container-based virtualization for applications has gained immense popularity thanks to the success of technologies like Docker. Container management is one of the key challenges of adopting this technology. As a result, management middleware like Kubernetes, Mesos, etc., are witnessing widespread adoption in the industry today. While Containers as a technology have reached an acceptable level of maturity, we see today that most of the challenges hindering the full-scale adoption of this technology lies in the limitation of the existing middleware managing containerized workloads. Problems around scalability, security, high-availability, disaster recovery, and compliance are still active research areas that require innovative solutions.

The aim of this workshop is to shed light on the main challenges and solutions of running containerized workloads in clustered environments.

The sixth workshop on container technologies and container clouds solicits contributions in this area from researchers and practitioners in both the academia and industry. The workshop welcomes submissions describing unpublished research, position papers as well as deployment experiences on various topics related to containers as outlined below:

Topics of interested include but are not limited to:

• Analyzing Scalability and performance of container management middleware (K8s, Mesos, Swarm...)

• Use cases for containers in Machine Learning workloads

• Security, isolation and performance of containers in shared environments

• Network architectures for multi-host container deployments

• Orchestration models for cloud scale deployments

• High availability systems for containerized workloads

• Leveraging hardware support for containers and containerized workloads

• Migrating and optimizing traditional workloads for containers

• Operational issues surrounding management of large clusters of containers

• Container use cases and challenges for HPC, Big Data and IoT applications

• Leveraging cognitive techniques for container management

• Performance enhancement of containers

• Use cases of using containers such as serverless computing and PaaS

• Comparative studies between different middleware for managing containers

• Comparative studies between containers, uni-kernels, and any other virtualization technologies

• Other topics relevant to containers management

Submissions:

Accepted papers should be no longer than 6 pages in the standard ACM format. Note that at least one author on each accepted workshop paper must hold a full pre-conference registration. As in previous years, the Middleware conference will provide companion proceedings including all workshop papers, which will be available in the ACM Digital Library. This is subject to the availability of camera-ready version by October 15, 2020. Please upload your papers in PDF form to hotCRP.

Submissions will be judged on novelty, relevance, clarity of presentation, and correctness. Authors of accepted submissions are required to present their work. Accepted papers and abstracts will be made available in the conference website at least one week before the workshop, so that the participants can come prepared having read the papers. Accepted submissions will be published via ACM Digital Library

• For the authors of accepted papers, you will be contacted by the main conference with a specific URL for your camera-ready submission.

Important dates:

• Paper submissions: September 21, 2020 - Extended: September 28, 2020

• Notification of acceptance: October 5, 2020

• Camera ready version: October 15, 2020

Presentations:

This will be a virtual event. Presentation will be made over zoom. The main conference, Middleware, will offer support with:

– Zoom for live face-to-face presentations and discussions;

– Slack channels for announcements and text-based discussions;

The authors are highly encourage to record their 15 minutes presentations, just in case they face technical difficulties during the workshop. Note that the authors have to do the presentation, and answer the questions live during Q&A. Instructions for uploading the recorded presentation can be found here.

The workshop will have a concise format for the presentations with a 15 minute slot for the talk, followed by 5 minutes of Q&A.

Note: at least one author must register for the conference workshop here. For all other participants, the registration is free here. Please register to receive the Zoom and slack information.

Program: December 7, 1:00 pm to 4:15 pm (Eastern Time, GMT -5)

Session 1: 1:00 pm - 2:00 pm

1:00 pm - 1:05 pm Workshop opening

1:05 pm - 1:30 pm Keynote: Multi-Cloud Management: Challenges and Opportunities

1:30 pm - 1:50 pm Understanding the performance of container execution environments

1:50 pm - 2:00 pm Short break

Session 2: 2:00 pm - 3:10 pm

2:00 pm - 2:20 pm Container Hardening Through Automated Seccomp Profiling

2:20 pm - 2:40 pm Filenail: Working with Incomplete Filesystem State

2:40 pm - 3:00 pm Flexible Migration in Blue-Green Deployments within a Fixed Cost

3:00 pm - 3:10 pm Short break

Session 3: 3:10 pm - 4:15 pm

3:10 pm - 3:30 pm Deepstitch: Deep Learning for Cross-Layer Stitching in Microservices

3:30 pm - 3:50 pm Heterogeneous MacroTasking (HeMT) for Parallel Processing in the Cloud

3:50 pm - 4:10 pm Cheetah: A fast unsupervised learning technique to provision next generation network services

4:10 pm - 4:15 pm Closing note

The entire conference program can be found here. Please note that on the conference website, the program is shown in Central European Time - Amsterdam. The time zone shown above is United States Eastern Time, since most of the papers came from North America. Check here to see the time in your timezone.

Keynote:

Title: Multi-Cloud Management: Challenges and Opportunities.

Abstract:

Studies show that more than 80% of the enterprises are interested in a multi-cloud approach, where their workloads are distributed across multiple clouds (private-private, private-public, public-public). A key enabler of the multi-cloud approach is standard management APIs and standardized packaging of the services. Kubernetes and Linux containers offer a solution for those requirements. However, the challenge remains in developing the middleware layer that glues the multi-cloud solution together. In this Talk we will present the challenges and the opportunities in multi-cloud management using Kubernetes and Linux containers.

Speaker's Bio:

Michael Elder is the Senior Distinguished Engineer for Red Hat Advanced Cluster Management for Kubernetes. His technical focus is enabling enterprises to deploy a common application programming model and operational model based on Kubernetes in hybrid architectures deployed across multicloud infrastructure. Prior to joining Red Hat, Michael led the launch of IBM Cloud Private and helped launch IBM Multicloud Manager, which both focused on enabling an open, hybrid platform based on Kubernetes. At IBM, he led DevOps solutions including IBM UrbanCode and IBM Bluemix Continuous Delivery. Through his leadership and execution, IBM was able to offer automated deployment capabilities for clients’ applications, into a variety of cloud delivery platforms. He most enjoys solving customer problems to enable IBM’s customers to provide better experiences for their customers.

Michael has delivered talks at IBM Think, O’Reilly Software Architecture, VMworld, IBM Interconnect, OpenStack Summit, SaltConf, ChefConf, and EclipseCon on topics ranging from his open source contributions to the Eclipse project to hybrid cloud deployment strategies and proof points. He has numerous awarded patents and has been honored with an IBM Corporate Award for technical achievement and three IBM Outstanding Technical Achievement awards. He is a co-author of Kubernetes in the Enterprise published by O’Reilly. Michael holds a B.S. in Computer Science from Furman University and an M.S. in Computer Science from the University of North Carolina-Chapel Hill.

Organization:

Workshop Chairs:

Ali Kanso – Microsoft

Seetharami R. Seelam – IBM T. J. Watson Research Center

Publicity Chair:

Chen Wang - IBM T. J. Watson Research Center

Program Committee Chair:

Abdelouahed Gherbi, Ecole de Technologie Superieure, Montreal, Canada

Program Committee:

Ricardo Koller, Google, United States of America

Abdelouahed Gherbi, Ecole de Technologie Superieure, Montreal, Canada

Abhishek Gupta, Facebook, United States of America

Ahmed Bali, Ecole de Technologie Superieure, Canada

Aleksander Slominsky, IBM Research, United States of America

Ali Kanso, Microsoft, United States of America

Chen Wang, IBM Research, United States of America

Guillaume Rosinosky, UCLouvain, Belgium

Jinho Hwang, IBM Research, United States of America

Manar Jammal, York University, Canada

Edilmo Palencia, Microsoft, United States of America

Parisa Heidari, Ericsson, Canada

Yogesh Barve, Vanderbilt University, United States of America

Nadjia Kara , Ecole de Technologie Superieure, Canada

Wubin Li, Ericsson, Canada

Workshop Recording: