Browse Prior Art Database

System and Method for Template Auto Scaling in Hybrid Cloud

IP.com Disclosure Number: IPCOM000247559D
Publication Date: 2016-Sep-17
Document File: 5 page(s) / 69K

Publishing Venue

The IP.com Prior Art Database

Abstract

In a hybrid cloud (including physical machines, virtual machines and linux containers) environment, the template means the image that will be used to create the physical machine (PM), the virtual machine (VM) or the container. When cloud user want to create an application with a large template, the cloud platform will pull the template from remote template repo if the template does not exist on local server. The end user need to wait a long time before their applications ready to use, and most of the time are spend in pulling the template from remote repo to local server. This innovation was trying to introduce a new dimension to “Auto Scaling” named as “Template Auto Scaling”. The “Template Auto Scaling” will help to manage the templates in cloud intelligently based on workload: pull template from remote repo and delete template from local server based on workload with some intelligent method. This can help speed up application deployment as well as application and infrastructure auto scaling.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 47% of the total text.

Page 01 of 5

System and Method for Template Auto Scaling in Hybrid Cloud

In a hybrid cloud (including physical machines, virtual machines and linux containers) environment such as OpenStack, CloudStack, Eucalyptus etc,the template means the image that will be used to create the physical machine (PM), the virtual machine (VM) or the container.

When cloud user want to create an application, the cloud platform will pull the template from remote template repo if the template does not exist on local server. The problem is that sometimes, the template is very big and it takes several minutes to download a template from repo, this caused end user need to wait a long time before their applications ready to use, and most of the time are spend in pulling the template from remote repo to local server.

This behavior also bring some issue to auto scaling, the current auto scaling in cloud is mainly referred to as application andinfrastructure resources auto scaling. But both application and infrastructure resources may need to pull a template remotely to local server to scale the application or the infrastructure, the remote pulling of template based on demand makes the application and infrastructure resources scaling not efficient as quite a lot of time are waste in pulling large templates.

There are some discussions for Mesos with docker image in MESOS-2824 , the issue is that a large image that will take a long time to initially fetch/hash/extract when the first container is provisioned. A proposed solution is support an operator endpoint so as to enable the operators can specify images to pre-fetch. This can indeed save sometime and can help speed up application deployment, but this brings a lot of manual work to maintain those images: which image should an operator pre-fetch, where should the image be pre-fetched to, when to delete the image if there are limited disk space on the server, all of those issuesare difficult to resolve.

This innovation was trying to introduce a new dimension to "Auto Scaling" named as "Template Auto Scaling". The "Template Auto Scaling" will help to manage the templates in cloud intelligently based on workload: pull template from remote repo and delete template from local server based on some intelligent method. This can indeed help speed up application deployment as well as application and infrastructure auto scaling.

Before go through details for this invention, we can take a look at five key words.

Threshold for Cold Template Threshold for Hot Template

Cold Template

Hot template:The template which is used widely in the cloud and need to scale up. The widely used means that the template "" Scale up a newtemplate means pull the template from remote template repo to a server which does not have such template.

Cold template:The template which does not use for a long time and need to scale down. Scale down a template means delete the template from aspecified server to free up some disk space so as to enable the server accep...