Friday, May 8, 2009

FEDERAL GOVERNMENT FLIPPING THE SWITCH

"Federal Web Portal Moves to Cloud Computing Platform" Read more on - http://www.govtech.com/gt/654240

This is another first, the US government itself moves on to the cloud; starting with its public information website USA.gov. This news breaks out a 'resounding yes' to CC and takes it closer to being a well accepted standard. While the driver may have been maintenance cost reduction by over 90%, I see this reflects the beginning of a cultural shift. Well, this sets ground for others to follow. A showcase for the ones who have been hesitant until now.

Another interesting fact that falls out is that this US government contract did not go one of the bigger ISVs like (IBM, Accenture, etc) but a much lesser known company Terremark. Repeating my comment form an earlier post, this year is going to be important for companies / professionals who are in anyway related to the cloud, while for the others this year will present opportunities that they will not be able to ignore. So it takes a totally different breed of companies that will partner and gain from this disruption. Most notably the ones who have been able to identify this shift and have been flexible enough (both culturally and strategically) to adapt accordingly.

This is not going to be a usual battle but a fight of standards.

Thursday, May 7, 2009

CLICKS AND MORTAR

Let me start by repeating myself. SaaS / Cloud computing is still in the adoption phase and as with any other new technology there is a list of issues that need to be addressed first. Even so, I feel more positive writing this post than writing the previous one. Perhaps my personal biases towards a more virtual environment.

In a more generic view, CC (cloud computing) is not really a new invention, but a phenomena sequential to of the desire of making information more accessible to the relevant mass; economically and logistically by leveraging our ever increasing computing power. Outsourcing information services to the cloud only makes them more scalable, flexible and affordable while keeping them much simpler, and I would argue, equally secure. No I am not being much ambitious here. And here is why:

First off, by embracing CC, all enterprises big and small evergreen their technology and IT infrastructure. The coming year is expected to be significant for the enterprises and professionals who are in anyway associated with CC, while for the rest it is going to be a year where they can certainly standby but surely cannot ignore this shift. My empathies.

Like most disruptive changes, this is going to take most of us (read them) by surprise. The shift has gathered pace and the ones to adopt it early are at an advantage over the others who have not. Whether it is about technology ever-greening or enjoying reduced costs and hitting a higher return on capital.

For starters, any CC app is much easier and faster to implement and equally easy to modify per the changing business environments. Most would agree (and feel sorry) that the traditional IT development and deployments are much time consuming and mostly a continuous affair inducing a need to maintain a complete IT department in-house.

While some of the (currently) popular traditional applications are much more comprehensive than their corresponding CC apps, they take a severe beating when it comes to utility of their comprehensive features. Most enterprises end up using a small percentage of these suites but continue to pay for the rest of the shelf-ware. With CC you pay only for what you use and there is mostly no shelf-ware. Plus CC is mostly elastic and can expand as and when required. This, by making the IT dollar go farther than ever, makes more economic sense for the huge dollar outlays on IT spending. Again, making it possible to reduce the IT staff and their incidental expenses; and replacing large up-front financial outlays with smaller, streaming yearly payments. Easily, Microsoft applications - Office and Exchange - cost 5 - 6 times more than Google Apps on a 3 year Total Cost of Ownership (TCO) basis. (Read - http://blogs.zdnet.com/storage/?p=472&tag=nl.e550 )

Perhaps the most reassuring fact is that the IT departments do not have worry about managing irregular updates, versions, fixing bugs and patches. With CC the upgrades are seamless and are always done by the service provider, and most importantly, are free. Further, by virtue of being hosted, the providers are able to offer the latest functionalities. For the providers themselves it is then about managing only one update / one release / one bug fix across the shared infrastructure and so they are able to sponsor it. They are equally incentivized to keeping their systems up to date with the latest capabilities in the market.

Now let’s have a quick look at performance. Performance and availability are perhaps the most notable and I'd say valid concerns yet. How dependable the cloud services really are? Will my critical applications be there when I need them? Well, any application whether installed or hosted is only as strong as the weakest link. No one of them is more vulnerable than the other. In that again CC takes the lead. By being on CC the companies are ensured of a world class service level which is not the case with smaller companies with limited IT infrastructure. The only weak link that may cause an interruption to business is the lack of stable bandwidth. This again is not totally unique to CC. The leading providers of CC are able to offer very transparent and credible service levels which are much in line with and sometimes beyond the requirements of companies.

Performance is also defined by how CC helps companies leverage a unified platform. It enables standardize IT processes for MNC that are spread across geographies and provide a collaborative work environment which was missing until now. To begin with isn't it about being able to access and share information easily, affordably and securely?

I think it's reasonably clear that the 1980's traditional distributed computing is passé and is beginning to crumble under its own weight. Cloud Computing is the way of the future.

Monday, February 23, 2009

IS IT REALLY BRICKS AND MORTAR?

The question is, "How willing are enterprises to migrate to the cloud?"

While most have clung on the conventional client server architecture, many entrepreneurs today have embraced the cloud. The critics like to believe that only 16% (the innovators and the early adopters in the curve) have been converted yet with little progress forecasted. I would like to argue that this is a good start for any form of 'disruption'.

The business applications was first moved to SaaS / cloud by companies like Salesforce.com and WebEx (Cisco) but the other bigger players like IBM, Oracle and SAP have caught up since. In the last three years of working on this platform I have been able to gather much insight about this issue directly from the horse’s mouth. Here are some reasons why companies will be slow or perhaps not adopt the platform.

SaaS / Cloud computing is still in the adoption phase and as with any other new technology there is a list of issues that need to be addressed first.

Most importantly and indeed the most feared reason is that the companies do not want their proprietary information to sit on someone else’s computers which are outside their firewalls. Especially when these servers are a part of a shared infrastructure that is not under the control of their IT departments. They store data for other companies too and I do now want my information to be available to the masses. The feeling is that data is more vulnerable because in a way it is on the web and hence accessible to the masses. Also, in some cases regulatory requirements prevent such data insecurity.

Another interesting point is that almost all of the providers have a licence fees based subscription to their service. Their subscriptions are mostly contracted for a period of one year (rarely longer in case the client is willing to sign up for longer) with no binding on the future contracts. One fears what the subscriptions are going to cost a few years down the line. Once companies customize the platform and migrate their systems and data, they tend to get used to working like that. But then they suspect that the provider may escalate the license fees in the future. It is like the utilities. You do not own the infrastructure but you get to use them for a rent. You really do not control how the rents look like in future. Can you possibly part with these services even if the subscriptions are doubled next year?

It is also widely said that the SaaS / Cloud in not (yet) equipped for the mainstream companies. One, it still has to prove itself, and two; the current applications do not support the million types of transactions that are captured by the conventional systems. There isn't enough functionality for companies to migrate to the cloud with their complex systems. Further there is a very limited amount of customizations that can be done on the cloud. If it does not fit our business processes that well then it is all in retrospect.

Even if companies were to migrate their systems to SaaS 'in part' to start with, there stands the challenge of integrating it with their current systems that are not migrated. The very purpose of being on a collaborative platform is defeated right there. IT departments are not willing to re-write a major part of the code in their current applications to make it compatible. And then there are partners that we work with where we have no control over their IT policies. If I implement it chances are that I will be the only one implementing it. And that immediately makes me isolated from the partner's who are not willing to migrate. What pinches the most it that a roll-back (if any) in the future is going to be even more painful. If we ever choose to go back to our conventional client-server architecture we will need to undo the entire exercise again and who know what problems may arise then. The other trend that is seen that some companies have moved on from one cloud to another. Even in such cases migrations isn't really straight forward. One could does not accept the data from another unless it is transformed. The IT departments aren't really very familiar with such platforms and migrations are an expensive part of the exercise.

Let’s talk a little about the performance. Since the application sits on the cloud, the performance is largely dictated by the speed of the internet available to the users. Companies adopting the cloud will have to accept the latency that comes with the use of internet. The current market segment consists of majorly SMBs where a lot of them aren't able to provide high speed broadband to all the users. At the same time, who guarantees the availability of the network? Obviously it is out of control of the provider but it still remains the backbone.

Finally, there are issues related to bureaucracy. The bigger the company the longer it will take them to implement a change of this size. Some companies that I worked with had a customized standalone application for each department. That makes it about thousands of applications in companies as big as Accenture and IBM. At the smaller companies on the other hand, it sounds like axing your own feet. If they outsource their IT infrastructure to SaaS provides what’s going to happen to the IT workforce?

So is it really The Next Big Thing in computing? Not qualified until we have looked at the other side of the coin. I will do so in my next post - 'Clicks and Mortar'