SBIR Phase I: Intelligent, real-time migration of software containers to optimize cloud computing resources

小型企业创新研究(SBIR)第一阶段项目的更广泛影响是提高计算云资源的利用率，以最大限度地减少浪费。企业越来越多地在云中运行IT基础设施和运营，以降低服务器费用。然而，从主要提供商那里租赁同等的云空间可能会让公司损失高达收入的50%。这在一定程度上是因为企业必须过度配置，以适应服务器使用的潜在激增，并确保无法容忍任何停机的应用程序不会中断。这种过度配置每年导致高达60%的未使用/空闲云空间和估计140亿美元的浪费支出。该项目将使用户能够通过利用现货市场上提供的大幅折扣的服务器空间，将其云支出减少高达90%，并将进一步使用户能够通过整合虚拟机来优化其工作负载，进一步降低成本。这个小型企业创新研究(SBIR)第一阶段项目将专注于开发两个基于智能自动化的新型解决方案。第一种技术将通过在折扣的Spot实例上生成容器，并根据可用性和价格在此类实例之间动态重新定位容器，显著提高关键应用程序的折扣虚拟机(VM)实例(Spot实例)的可用性。第二种技术将通过在空闲期间将空闲容器打包到少量VM上来解决云中的空闲资源浪费问题，从而最大限度地减少活动VM的数量，从而降低保持服务在线的成本。当工作负载增加时，该技术会将容器重新部署到不同的虚拟机上，而不会造成任何服务中断。为了确定可行性，将为每个要素设计和开发一个原型。目标包括：1)优化网络和存储，以支持公司的专利容器架构，以安全地隔离云中的容器；2)为两个建议的解决方案开发资源调度器3)开发用于工作负载监控和配置的用户界面。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

The broader impact of this Small Business Innovation Research (SBIR) Phase I project is to improve usage of computing cloud resources to minimize waste. Increasingly, businesses are running IT infrastructure and operations in the cloud to reduce server expenses. However, leasing equivalent cloud space from major providers can cost companies as much as 50% of their revenue. This is in part because businesses must overprovision to accommodate potential surges in server use and to ensure that applications that cannot tolerate any downtime are not interrupted. This overprovisioning causes as much as 60% of unused/idle cloud space and an estimated $14 Billion in wasted spending annually. This project will enable users to reduce their cloud spend by up to 90% by taking advantage of deeply discounted server space available in spot markets, and it will further enable users to optimize their workloads by consolidating virtual machines, further reducing cost. This Small Business Innovation Research (SBIR) Phase I project will focus on the development of two novel intelligent-automation-based solutions. The first technology will dramatically improve the usability of discounted virtual machine (VM) instances (Spot Instances) for critical applications by spawning containers on discounted Spot Instances and dynamically relocating containers between such instances based on availability and price. The second technology will address idle resource waste in the cloud by packing idle containers onto a small number of VMs during the idle period, minimizing the number of active VMs and thus reducing the cost of keeping services online. When the workload increases, the technology will relocate containers onto different VMs, without any service interruption. To establish feasibility, a prototype for each element will be designed and developed. Objectives include: 1) optimize network and storage for supporting the company’s patented container architecture for secure isolation of containers in the cloud; 2) develop a resource scheduler for both proposed solutions 3) develop a user interface for workload monitoring and configurations.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.