Cloud computing nowadays is not an emerging topic, and virtualization is an indispensable technology to expedite cloud computing to become the next sign of the coming Internet revolution. In real life, scientists never stop at exploring the possibilities from such technology by investigating millions of experiments and applications to enhance the quality of virtual services. However, isolated construction for the virtual machine doesn’t save the technology from unwanted data volumes or insensitive processing time. Containers are created to address such problems, by distributing applications without initiating the entire virtual machine. Docker, as an important player in this game, is an open-source application of the container family. The management tool from Docker containers, Swamskit, does not take heterogeneities in either virtualized containers or physical nodes. There are different nodes in the cluster, and each node is different in configurations, resource availability, or concerning resource, etc. Furthermore, the requirements initiated by different services change all the time. The demand might be CPU-intensive (e.g. Clustering services) and also memory-intensive (e.g. Web services), or completely at the opposite. In this paper, we focus on exploring the Docker container cluster and designing, DRAPS, a resource-aware placement scheme, to improve the system performance in a heterogeneous cluster.

In the past decade, we have witnessed a dramatically increasing volume of data collected from varied sources. The explosion of data has transformed the world as more information is available for collection and analysis than ever before. To maximize the utilization, various machine and deep learning models have been developed, e.g. CNN [1] and RNN [2], to study data and extract valuable information from different perspectives. While data-driven applications improve countless products, training models for hyperparameter tuning are still time-consuming and resource-intensive. Cloud computing provides infrastructure support for the training of deep learning applications. The cloud service providers, such as Amazon Web Services [3], create an isolated virtual environment (virtual machines and containers) for clients, who share physical resources, e.g., CPU and memory. On the cloud, resource management schemes are implemented to enable better sharing among users and boost the system-wide performance. However, general scheduling approaches, such as spread priority and balanced resource schedulers, do not work well with deep learning workloads. In this project, we propose SpeCon, a novel container scheduler that is optimized for shortlived deep learning applications. Based on virtualized containers, such as Kubernetes [4] and Docker [5], SpeCon analyzes the common characteristics of training processes. We design a suite of algorithms to monitor the progress of the training and speculatively migrate the slow-growing models to release resources for fast-growing ones. Specifically, the extensive experiments demonstrate that SpeCon improves the completion time of an individual job by up to 41.5%, 14.8% system-wide and 24.7% in terms of makespan.

In the last decades, we have witnessed a spectacular information explosion over the Internet. Millions of users are consuming the Internet through various services, such as mobile applications, and online games. The service providers, at the back-end side, are supported by state-of-art infrastructures. Targeting on providing the services at scale, virtualization is one of the emerging technologies used in data centers and cloud environments to improve the quality of services. In this project, we aim to develop a dynamic resource management scheme based on virtual containers. It collects the runtime job progress from the running tasks and allocates the resources dynamically to improve the overall system performance.

Containers are gaining popularity in life science research as they provide a solution for encompassing dependencies of provisioned tools, simplify software installations for end users and offer a form of isolation between processes. Scientific workflows are ideal for chaining containers into data analysis pipelines to aid in creating reproducible analyses. In this manuscript we review a number of approaches to using containers as implemented in the workflow tools Nextflow, Galaxy, Pachyderm, Argo, Kubeflow, Luigi and SciPipe, when deployed in cloud environments. A particular focus is placed on the workflow tool’s interaction with the Kubernetes container orchestration framework.

Kubernetes (K8s) defines standardized APIs for container-based cluster orchestration so it becomes possible for application managers to deploy their applications in a unified manner across different cloud providers. A practical problem is however feature incompatibility between different K8s vendors, who offer commercial K8s products based on the open-source K8s distribution. A large number of documented features in this open-source distribution are optional features that are turned off by default, but can be activated by setting specific combinations of parameters and plug-in components in configuration manifests for the K8s control plane and worker node agents. However, none of these configuration manifests are standardized, giving K8s vendors the freedom to hide the manifests behind a single, more restricted, and proprietary customization interface. Therefore some optional K8s features cannot be activated consistently across K8s vendors and applications that require these features cannot be run on those vendors. In this paper we present a unified, vendor-agnostic feature management approach that bypasses the proprietary customization interface of K8s vendors in order to consistently activate optional K8s features across a federation of clusters hosted by different Kubernetes vendors. We describe vendor-agnostic reconfiguration tactics that are already applied in industry and cover a wide range of optional K8s features. Based on these tactics, we design and implement an autonomic controller for declarative feature compatibility management across a cluster federation. We found that the features configured through our vendor-agnostic approach have no impact on application performance when compared with a cluster where the features are configured using the configuration manifests of the open-source K8s distribution. Moreover, the maximum time to complete reconfiguration of a single feature is within 100 seconds, which is 6 times faster than using proprietary customization interfaces of mainstream K8s vendors such as Google Kubernetes Engine. However, there is a non-negligible disruption to running applications when performing the reconfiguration; this disruption impact does not appear using the proprietary customization methods of the K8s vendors. Therefore, our approach is best applied in the following three use cases: (i) when starting up new K8s clusters, (ii) when optional K8s features of existing clusters must be activated as quickly as possibly and temporary disruption to running applications can be tolerated or (iii) when proprietary customization interfaces do not allow to activate the desired optional feature.

This paper presents a groundbreaking approach to real-time image processing in electronic component assembly, enhancing quality control in manufacturing. By capturing images from pick and place machines between component pickup and mounting, defects are identified and addressed in-line, significantly reducing the likelihood of defective products. Leveraging fast network protocols like gRPC and orchestration with Kubernetes, along with C++ programming and TensorFlow, this method achieves an average turnaround time of less than 5 milliseconds. Tested on 20 live production machines, it ensures compliance with IPC-A-610, and IPC-STD-J-001 standards while optimizing production efficiency and reliability.