Where is the Cloud?

A chat with Zia Khan, founder of Operation Badar
Believe it or not, this Cloud that everyone talks about, has been around forever. Like most things, technology advancement, intervention coupled with professional innovation has been able to recycle concepts and repackage them into something more practical and significant.

Technology, in its essence, is about making processes more efficient, optimizing resources and maximizing the bottom line. Organizations depend on IT departments and the CIO to opt for solutions which help to achieve this result. Thin Client Computing is all about reducing the burden in the real world and load balancing it with the support extended by the virtual world. Better server management, virtualization, best practices for disaster management along with faster, more redundant internet connectivity enables a more resilient environment for companies to work in.

Ziaullah Khan, the founder of Operation Badar explains. “Think back to the fundamentals. The concepts have remained the same. People started off working in isolation. Take the example of electricity or utilities where factories have their own, individual power generation plant.” Zia continues the explanation that the individual power generation plants will become inefficient. “It isn’t going to run at 100% capacity all the time and regardless of the fact whether you have one light on or 100, you will run the plant 24 hours. Add to that the cost of maintenance, diesel and other expenses. This makes sense in the 1850’s in the US because there was no centralized power grid. The question of national grid comes along, where one power generator is installed and everyone consumes electricity from it.” And that shared consumption, gives rise to the concept of Load Balancing to give optimal utility.

“In the morning, perhaps offices need more electricity, but residential units need more at night. So you plan the grid to manage the load accordingly. The entire concept of Cloud Computing, is essentially built on the same premise.”

Turning Computing Power into Utility
You can argue that if the concept of the Cloud has been around for such a long time, why it wasn’t being harped about some five years ago. The answer, Zia explains, is as simple as the solution. “Cloud Computing requires economies of scale to work. The demand has to be there for it to make real sense. This was not the case even 5 years ago because everything was a lot more expensive than it is today. The cost for storage space, computing power and the CPU have become relatively less cost-intensive because the demand is so great. And as these costs are decreasing, the computing power, which is simple evolution, has increased.” For Microsoft’s legacy model of business, this cost-decrease-demand-increase is going to brew up a huge storm.

So, for the average CIO who has to make a decision about his own technology setup, he also needs to know what is going to happen in the next year. “Everyone can sell your pretty much the same equipment and configuration for relatively the same cost. But what makes you go with one vendor over another, is the Quality of Service or the Service Level Agreement (SLA) you will sign with them. The market becomes a more cut-throat now than ever before.”

The Cloud as a Service
That’s basically the whole idea. Companies are going to simply rent the servers from Microsoft, Amazon, Google, and other cloud-based computing providers. “You rent the space and computing power and from these cloud computing providers and pay depending on what you use. Even now, most of the companies who have their own servers, end up utilizing less than 20% of the power and capacity. Because remember, when you have your own server, you have to calculate your peak utilization and work around that.”

Just like the power grid example he started off with, Zia explains that the biggest problem with having your own servers is that you will always remain at 10% to 20% utilization capacity.

Take the example of Amazon, an e-retailer with a huge customer base, began selling books and now sells virtually everything. Some five years ago, Amazon encountered a problem. “The issue was that for most of the time, their equipment for most of the time was not used optimally. During Christmas, for example, they have high demand whereas the rest of the year, their servers are relatively idle. That was when their CIO decided to rent out the additional capacity.” And because there was just so much capacity and power there, it chucked out a revolution in the cloud space.

He goes onto explain, “It’s all about solving problems. Let more people use your computer without disturbing one another. That’s where server virtualization comes into the picture. In a virtualized environment, you can run multiple instances of different operating systems and configurations, making it think that it has direct access to the hardware.” In the virtualized environment, every operating system functions in isolation of the others. Therefore this virtualization layer sits between the software and the hardware. “So Amazon starts to rent out their space and begins to launch services. One of their services called Amazon Elastic Compute Cloud, or EC2, is a web service that provides resizable computing capacity in the cloud, and another one called Simple Storage Solutions or S3.”

The problem with Amazon, according to Zia, is that they are not software providers. They have taken a million normal PC’s servers running Linux, UNIX or Windows and plugged them into single network and any member of the cloud, can have access to them. “The problem is if I can rent one processor from them and my load increases, I would be sitting there writing the Load Balancing application myself. Anything else I need, I would have to do it myself because the host can’t provide it. And that brings us to different categories of cloud computing providers.”

Amazon is what is known as an Infrastructure As A Service (IAAS). “And let’s take a step away from Amazon and look at Google’s model for a second.”

Google launched many products and services under the banner of Software As A Service (SAAS). But if you look at their primary model, it’s interesting to see how the search engine, which is what their primary function is, made it possible for SAAS to go even one step further. Take a look.

When you type in a search query on Google, explains Zia, the search yields you a result in a matter of seconds. But it just doesn’t seem possible that it crawled through billions of pages of information to get you what you needed, especially when there are at least a few million other users searching for at least a few million diverse tags. Google has a server form already installed whereby simultaneously running your search query against its algorithms across hundreds of servers.

“So Google looked at the services and IAAS model that Amazon was launching and realized they could do something similar, but scale it to a larger size.” Google had more servers than Amazon and was already making a lot of money, whereby their profit margins were already close to 60%. “So they were liquid and had the size. Both things needed to put Microsoft down. Google knows that in the long run, Microsoft will give them a run for their money – Microsoft struggles when it comes to the search algorithm that gives Google it’s edge, but it will only be a matter of time before they catch up. You see, says Zia, Microsoft has the platform, browsers and operating system while companies are using their servers. “While Microsoft is strong in the enterprise space, Google is very strong in the consumer space. The motive for Google to step up their game for SAAS and convert it towards a Platform As A Service (PAAS) was tremendous.”

About 5 months ago, Google introduced the ‘Google App Engine’. “This is essentially a python-based system with a Java-based front end. This is the key to the PAAS as opposed to Amazon’s SAAS. “Platform As a Service provides you with the relevant tools so you now have load balancing, for example, taking place on the server automatically. If your site or service is running on Google’s App Engine, and the load increases, the servers will scale automatically; they will adjust given the number of users and the number of deployed machines and you get charged only for the time this increase takes place. Otherwise, you continue to pay your regular rental fee.

So Much Competition – What Next?
That’s the question everyone is asking. When Amazon got into this space and launched their services, Microsoft realized the threat. They brought the likes of Dave Cutler Sr., the main architect behind Windows NT, along with others to plan how to get Microsoft ahead in the game.

As the technology decision-maker in your organization, you need to know how one company differs from another in their hosting or contact solutions for you. The cloud is a big place and if you don’t know what the background on some of these service providers is, you are bound to get lost in unmanageable territory.

Next month, we’ll talk about how Microsoft’s response to Amazon and Google has been with Azure and how the Cloud is a viable option for companies in Pakistan.

About the Author
Zia Khan is the CEO of Operation Badar and has more than 12 years of experience in software architecture, design, development, implementation, and integration. He is also a Microsoft MVP. You can contact him through operationbadar.net