[Paper Review] Cloud Computing -- Everything As A Service

Michael Howard|arXiv (Cornell University)|May 25, 2022

Cloud Computing and Resource Management4 citations

TL;DR

This paper surveys cloud computing orchestration and management frameworks, proposing a vendor-agnostic approach using Tekton for CI/CD pipelines, CloudCAMP for infrastructure modeling, and Kubernetes for container orchestration. It demonstrates secure runtime via Asylo's Trusted Execution Environment, and introduces specialized standards like TOSCA4QC for quantum workloads and GeneSIS for IoT systems, enabling scalable, secure, and portable cloud deployments across diverse application domains.

ABSTRACT

Compute infrastructure hosted by a cloud provider allows an application to scale without limit. The application developer no longer needs to worry about the up-front investment in a server farm provisioned for a worst-case load scenario. However, managing cloud deployments requires a sophisticated framework that can autoscale the infrastructure and guarantee the up-time of running container images. This paper surveys existing research addressing the management and orchestration of cloud deployments as well as the modelling framework to abstract away the low-level details of the host infrastructure. We investigate blockchain distributed ledgers, quantum computing and Internet of Things application stacks to show how they can utilize cloud deployments.

Motivation & Objective

• To address the challenges of managing dynamic, on-demand cloud infrastructure across heterogeneous environments.
• To reduce vendor lock-in in CI/CD pipelines through vendor-agnostic frameworks like Tekton.
• To enhance runtime security of cloud-deployed applications using Trusted Execution Environments (TEEs).
• To enable portable, model-driven deployment of cloud and edge systems through frameworks like CloudCAMP and GeneSIS.
• To extend orchestration standards for emerging workloads such as quantum computing and IoT systems.

Proposed method

• Uses Tekton to define and automate CI/CD pipelines as declarative, containerized tasks within Kubernetes clusters.
• Employs the Asylo framework to wrap application binaries with Intel SGX-based encryption, enabling secure execution in TEEs.
• Applies the CloudCAMP framework to model cloud infrastructure abstractly, independent of underlying cloud providers.
• Compares Kubernetes and Docker Swarm as container orchestration solutions, emphasizing Kubernetes' flexibility and feature set.
• Introduces TOSCA4QC, a new extension to TOSCA, to model and automate quantum-compute application deployments in the cloud.
• Develops the GeneSIS framework using a domain-specific modeling language and execution engine to enable continuous orchestration of heterogeneous IoT systems with limited or no internet access.

Experimental results

Research questions

• RQ1 How can cloud application deployment be automated and decoupled from specific cloud providers?
• RQ2 What mechanisms enable secure runtime execution of containerized applications in untrusted cloud environments?
• RQ3 How can infrastructure provisioning and orchestration be abstracted to reduce cloud provider dependency?
• RQ4 What modeling standards are needed to support emerging workloads like quantum computing and IoT in the cloud?
• RQ5 How can continuous deployment and dynamic adaptation be achieved in resource-constrained IoT environments?

Key findings

• Using Tekton for CI/CD pipelines enables vendor-agnostic automation, avoiding the need to reconfigure pipelines when switching cloud providers.
• The Asylo framework increases build times by a factor of 65× due to TEE integration, but provides strong runtime security by encrypting code in memory via Intel SGX.
• CloudCAMP enables high-level, cloud-agnostic modeling of infrastructure, abstracting low-level provisioning details.
• Kubernetes introduces runtime overhead due to its rich feature set, but remains the dominant orchestration platform for containerized workloads.
• TOSCA4QC successfully extends TOSCA to model quantum-compute applications, with case studies proving its feasibility for cloud-based quantum workloads.
• GeneSIS enables platform-independent, continuous orchestration and deployment of Smart IoT Systems across cloud, edge, and IoT devices, even in environments with intermittent or no internet connectivity.