Virtual Desktop Infrastructure: Sharing is Caring!

(October 2010)

There’s a buzz word doing the rounds these days. They call it VDI in short, which expands to Virtual Desktop Infrastructure. 

Desktops have historically been a physical entity; then what is it with ‘virtual’ that we hear so often? Well, the crux of the concept around VDI is to move from ‘physical’ to ‘virtual’. Let’s look at it in detail.

What is VDI?

Imagine your desktop shrinking in size, and the CPU cabinet being removed from your desk. Stretch your imagination a bit further and visualize the OS and most of the local applications vanishing in a whisker. Does it make you cry, “My desktop crashed just now!” No, it has not. We just gave you a ‘dumb’ terminal that still has the ability to get you connected to Internet. And, we’ve smartly moved all the above ‘missing’ features to a data centre. It doesn’t hurt not to know where the data centre is physically located. That’s because you’ll be able to work on an image of the OS and images of all the applications that you’re so used to working on. While it does not violate the licensing agreements with vendors, this model allows you to ‘relax’ in the sense that you will have limited local issues with your desktop. For your daily work, you will have to access (through secure authentication) various applications and storage units which are hosted in the data centre. 

Why VDI?

Well, I ask – why not? Migrating from the traditional desktop settings to VDI makes a strong business sense, and the icing on the cake is: it pays you back financially. Let’s see how.

Windows 7 upgrade

Consider an organization which has 50 thousand desktops, each of which needs an upgrade to Windows 7. It would take nothing short of months to test, integrate and install the new OS onto all the 50,000 desktops; not to mention the workforce required to support any future OS issues, post-implementation. What if the OS is made to disappear from all the above desktops, hosted in the data centre, and an image of the same made available to each of the desktop users? Further patches and updates would be facilitated from the single centralized image. And the same system hardware would be more than smart to support the new OS. Any desktop user would say, ‘wow’! 

Disaster recovery and Data loss mitigation

We’re fast moving towards a distributed workforce model. One efficient way to ensure data protection in this scenario is to control it centrally. In the VDI environment, all user and customer data is hosted in the data center, leading to faster data recovery and lesser service downtime. Also – in the event of temporary service disruption, support staff need not rush to desktop users. 

Users’ data protection

Think of systems whose access needs to be controlled and audited periodically; or whose data are too sensitive and confidential to move out of the data center. That is easily facilitated by VDI with authentication logs that track which user accesses which applications at what point in time. VDI beats the traditional desktop environment thumbs down in this regard.

Field workers

How often is it that you find salespersons collecting data from you and then rushing back to their offices to update the same on their internal applications? Even laptops don’t always save them as they have to face data loss, system crash and other local problems. VDI says, they don’t need fancy devices to work on, while they are in the ‘field’; instead provide them with the basic connectivity that would enable them to work seamlessly anytime anywhere.

Fast organizational growth

If yours is a fast-growing organization, chances are you see new employees getting onboarded quite often. Imagine a workplace that equips them with lower-cost thin clients or re-used legacy endpoints instead of fully loaded PCs. Such re-used endpoints may not have the capability to run the latest OS or new applications. By making new employees work on such systems, you may compromise on the ‘look’, but isn’t ‘less’ more beautiful at times than ‘more’, especially when your capability with lesser resources scales up manifold? 

In the end, users from across business lines and geographies may share the same hosting environment, but they would work on their designated profiles. And they care not to poke into others’ domains or profiles.

Someone rightly said, sharing is caring. 

_________________________________

Application Virtualization: from Deployment to Delivery

(written in November 2010)
 
We live in a world that is not the world that we were born in. They say, change is the only thing that does not change. How true! Our traditional world essentially was made up of 'physical' substances and concepts. While most of the substances continue to be 'physical', the concepts around these sure are going the 'virtual' way.

We are virtualizing desktops, doing away with local operating systems, looking for data centers to host our resources. And what not!

Let us see how Jack - a user in the physical world - uses various Applications on his system. Except for a few web-based hosted applications, Jack would need to deploy and install applications 'locally' on his system. He may not need to access all the installed applications at all times, but he would not want to uninstall any application on his own, unless being directed to do so by the organization. In this model, he acts as a part time system administrator for himself, while there are dedicated teams of expert administrators and application support professionals. I am not saying this has not been a successful model of managing applications on end user systems; I am only asking a question: "Can there be an alternate approach of addressing the obvious pain areas in the traditional system, while ensuring a uniform standardized user experience?" I would attempt to answer in brief: "Why not?"

An organization wanting to manage traditional applications must go through the cycle of testing, deploying, installing, patching, updating, and finally uninstalling. Today's multi-hardware multi-OS platforms complicate this further, with a dedicated bunch of support staff spending tangible resources to manage such application lifecycles. Add to this the numerous challenges with consistently deploying applications on end user systems. What it translates into is an increasing number of organizations looking around for a simpler way of handling application management from end user perspective. 

Imagine Jack not requiring to deploy any applications at all, on his system, rather being able to work on streamed instances of operating systems and applications on demand. An application repository is created, applications profiled using an OS environment, profiled applications are encapsulated and stored in the secure repository. When Jack needs to work on or access a certain application, he places a request, which gets picked up by a virtualizer, that uses its 'brain' to ascertain if Jack's system is compatible to 'run' the requested application.

If 'yes', the requested application is packaged (with the required components) and streamed from the repository to the end system. In this case, the application uses the local system's resources for running and execution. If Jack's system is not compatible, the requested application is streamed on to a hosting environment in the data center, which is then made available to him. The application is executed and run on the server in the hosting environment.

In both the instances, Jack does not have to worry about anything but working on the applications alone. Any subsequent patches or updates to the applications are performed at the 'source' level, allowing one-time fix for all instances of the requested application at one point of time, and letting end users work on their core business areas. Let us not be surprised if Jack indeed becomes the master of all trades! Organizations can then afford to reduce their IT support staff strength. Jack also has the liberty of working 'offline' on the streamed instance of a requested application, while 'live' updates are being applied to the application at the source. Once Jack is able to again work on the online instance, the recent updates are automatically made available to him. Quite fittingly, this entire feature is known as Application Virtualization.

As is evident from the above, applications are no longer required to be 'deployed'; rather they are 'delivered' on demand, to end users.

Let us now take a deeper look at the concerns with traditional deployment-based application management.

·         Increasing complexity of catering to user demands of real time access to an   ever-changing set of applications; anytime, anywhere.

·         With organizations opting for newer operating systems, e.g. Windows 7, a lot of existing applications are becoming incompatible. This calls for replacement of legacy applications, leading to a big hole in an organization's pocket.

·         Repeat telecast of IT efforts due to endless lifecycle of updates, patches and security fixes.

·         Exposure due to poor application security and control.

·         High cost of overall application management.

·         High cost of IT support staff.

In short, to say the least, Application Virtualization is the way forward.