Welcome!

Machine Learning Authors: Pat Romanski, Yeshim Deniz, Liz McMillan, Elizabeth White, Zakia Bouachraoui

Related Topics: Machine Learning , IBM Cloud, @CloudExpo

Machine Learning : Blog Post

Cloud Computing and Virtual Images

A look at several different cloud computing solutions will reveal a technological enabler present in almost each one

Cloud Expo on Ulitzer

A look at several different cloud computing solutions will reveal a technological enabler present in almost each one. The enabler I'm talking about is the use of virtual images. I cannot think of many, if any, cloud computing solutions that provide software elements (i.e. more than just servers, storage, memory, etc.) that do not use virtual images in some form or fashion.

Of course, one of the reasons virtual images form the backbone of many cloud solutions is obvious. Virtual images provide the benefits of server virtualization. We can activate many virtual images on the same physical machine, thus allowing us to achieve multi-tenancy (multiple operating systems and software stacks installed on the same physical machine). Besides driving higher hardware utilization rates, it also provides us the capability to run heterogeneous software environments on the same piece of hardware. This both enables and encourages the creation of a shared pool of compute resources which is a key characteristic of cloud computing environments.

Server virtualization may be the first thing that comes to mind when we think about virtual images, but at least in the context of cloud computing, I do not believe this is the most important benefit. If we are looking at cloud computing as a means to quickly and consistently provision software environments, then I think virtual images provide us a capability and benefit more valuable than server virtualization. In this respect, virtual images provide us a medium through which we can templatize the configuration of our software environments.

Consider the case of a fairly basic application serving environment. In this environment, you are likely to install an operating system, application server, and probably some type of load balancing solution. Each of these typically requires a different piece of software, different installation procedures, and finally integration with the other components. Installing these into a typical environment, without the use of virtual images, means that you either have scripts that account for each different piece of software and finally integration of the different components, or it means that a person manually installs and integrates the pieces each time you need an environment. Either process can be time-consuming and costly to maintain over time.

Enter the use of virtual images. With a virtual image, you can install and integrate all three components ONE time, and then capture the resultant environment as a virtual image. At this point, when an application environment is needed, the virtual image can simply be activated on top of a hypervisor platform. The application environment is typically available in a much more timely fashion than if manually installed or integrated because the installation and configuration have already been captured in the virtual image.

From what I described above, you may have caught on to what would seem like a drawback of using virtual images to templatize software environments. Specifically, it may seem that you need a distinct virtual image for every unique configuration of your application environment. If this were the case, management of your virtual image library would soon become a nightmare and the resulting cost (in both resource and time) would likely outweigh the original benefits. However, thanks to a relatively new standards-based approach to virtual images, this is not necessarily a problem.

The standard I'm talking about is the Open Virtualization Format (OVF) standard which has been published by the Distributed Management Task Force (DMTF). According to the OVF standard, it "describes an open, secure, portable, efficient and extensible format for the packaging and distribution of software to be run in virtual machines." In particular to our discussion here, there is a part of the standard that describes the use of an ovf-env.xml file within the virtual image.

This file is essentially a key-value style XML file that describes desired aspects of the environment. Keys and values can be supplied during image activation, and configurations scripts that run during virtual image activation can read information from the file and react appropriately. Thus, instead of supplying N different virtual images for N different software environments, you can supply 1 virtual image and utilize the ovf-env.xml file in conjunction with configuration scripts within the image to produce N different environments. The use of this mechanism with virtual images delivers the capability to templatize software environments without sacrificing flexibility or encouraging unsustainable virtual image proliferation.

In WebSphere, we utilize the model outlined in the OVF standard when packaging our WebSphere Application Server Hypervisor Edition virtual images. This allows our WebSphere CloudBurst product to provision these images and create many different types of WebSphere Application Server environments from a single virtual image (read this article for more information). I expect the use of this standard and the mechanisms it provides will become pretty prevalent in the near future. Now if we could just get to the point where virtual disk formats are standardized, but that's an entirely different topic.

More Stories By Dustin Amrhein

Dustin Amrhein joined IBM as a member of the development team for WebSphere Application Server. While in that position, he worked on the development of Web services infrastructure and Web services programming models. In his current role, Dustin is a technical specialist for cloud, mobile, and data grid technology in IBM's WebSphere portfolio. He blogs at http://dustinamrhein.ulitzer.com. You can follow him on Twitter at http://twitter.com/damrhein.

CloudEXPO Stories
With more than 30 Kubernetes solutions in the marketplace, it's tempting to think Kubernetes and the vendor ecosystem has solved the problem of operationalizing containers at scale or of automatically managing the elasticity of the underlying infrastructure that these solutions need to be truly scalable. Far from it. There are at least six major pain points that companies experience when they try to deploy and run Kubernetes in their complex environments. In this presentation, the speaker will detail these pain points and explain how cloud can address them.
The deluge of IoT sensor data collected from connected devices and the powerful AI required to make that data actionable are giving rise to a hybrid ecosystem in which cloud, on-prem and edge processes become interweaved. Attendees will learn how emerging composable infrastructure solutions deliver the adaptive architecture needed to manage this new data reality. Machine learning algorithms can better anticipate data storms and automate resources to support surges, including fully scalable GPU-centric compute for the most data-intensive applications. Hyperconverged systems already in place can be revitalized with vendor-agnostic, PCIe-deployed, disaggregated approach to composable, maximizing the value of previous investments.
When building large, cloud-based applications that operate at a high scale, it's important to maintain a high availability and resilience to failures. In order to do that, you must be tolerant of failures, even in light of failures in other areas of your application. "Fly two mistakes high" is an old adage in the radio control airplane hobby. It means, fly high enough so that if you make a mistake, you can continue flying with room to still make mistakes. In his session at 18th Cloud Expo, Lee Atchison, Principal Cloud Architect and Advocate at New Relic, discussed how this same philosophy can be applied to highly scaled applications, and can dramatically increase your resilience to failure.
Machine learning has taken residence at our cities' cores and now we can finally have "smart cities." Cities are a collection of buildings made to provide the structure and safety necessary for people to function, create and survive. Buildings are a pool of ever-changing performance data from large automated systems such as heating and cooling to the people that live and work within them. Through machine learning, buildings can optimize performance, reduce costs, and improve occupant comfort by sharing information within the building and with outside city infrastructure via real time shared cloud capabilities.
As Cybric's Chief Technology Officer, Mike D. Kail is responsible for the strategic vision and technical direction of the platform. Prior to founding Cybric, Mike was Yahoo's CIO and SVP of Infrastructure, where he led the IT and Data Center functions for the company. He has more than 24 years of IT Operations experience with a focus on highly-scalable architectures.