Archive for the ‘Cloud Computing’ Category

Understanding the Cloud Computing Vendor Landscape

January 18, 2010

More than six months back, I had written about the tremendous amount of hype surrounding Cloud Computing. Then few months after that I took some clues from the Gartner Hype Cycle Report and said that this hype around Cloud Computing has reached the peak and now has started declining (Ref – “Beyond the hype – can some real work using Cloud Computing start now?“). IMO, real work or usage of any technology typically starts once the products shreds some of the hype or unrealistic expectations around it. As we work with our customers at GlobalLogic in consulting/helping them on various aspects of Cloud, I am now seeing the validation of my prediction.

However, now we have started seeing new challenge facing many who are taking or planning to take their first steps towards the implementation – figuring out which solution or technology does what and which solution to select by going through the mesh of vendors out there. The reason I say this is because Cloud Computing is a paradigm which does not suit itself to home development. Any solutions built up on or using Cloud would have to be built up by using or in partnership with the vendors out there. Deciphering the confusing marketing message sent out by various vendors about their capabilities and how these capabilities fits with the requirements in hand can be an overwhelming task by itself.

Considering the above challenge – recently I came across the vendor taxonomy created by Peter Laird which I thought was an impressive start in terms of laying the lay of the land in Cloud Computing.

Cloud Vendor Taxonomy - 2009

Cloud Vendor Taxonomy - 2009 - By Peter Laird. (Click on image for bigger view)

This blog post from Peter also describes what each of the category stands for. It is worth reading. Although based on our experience here some of the specific vendor classification could be debatable either way – however I am in agreement with the overall structure of the taxonomy. In our business at GlobalLogic in the Consumer-oriented applications side, we have explored solutions provided by the Public Cloud, Business User Platforms, Development Platforms along with Storage and Integration platforms. As we are enhancing our expertise in building applications/products on the Enterprise side – next steps for us would be to explore the Private Cloud side along with Billing and Integration side. Also my feeling is that Infrastructure-related technologies (as listed in the figure) would be mostly seamlessly packaged in by the Private Cloud Vendors.

In any regards, just wanted to share this pretty useful taxonomy created by Peter Laird to the readers of my blog. Thoughts and comments are welcome.

Beyond the hype – can some real work using Cloud Computing start now?

October 1, 2009

Few months back in the month of June, I had wailed against the increasing marketing hype and the resulting over-expectations set on Cloud Computing which IMO is a powerful concept but yet is still at its nascent stage. (read my blog post titled “Cloud Computing – Is this the case study for Marketing gone out of hand?“). Then in the very next month of July, Gartner came out with their much anticipated annual “Gartner’s Hype Cycle Special Report for 2009” which in some subtle way examines the maturity of emerging technologies.

(For those unfamiliar with what Hype Cycle means, you may want to refer to this Wikipedia description. Coined by Gartner, the Hype Cycle hypothesizes that any emerging technology typically goes through an initial period of over-enthusiasm and popularity which in some sense results in inflated expectations from the emerging technology. After this initial period, the hype around this emerging technology typically falls possibly because of  disenchantment. Few technologies recover from this phase of disenchantment with a practicality/sanity of usage taking over.)

As the 2009 Hype curve shows below – Gartner feels that Cloud Computing has now reached the peak of the hype cycle and based on their theory it many now be entering into the depths of disenchantment.

2009 - Gartner Hype Cycle

2009 - Gartner Hype Cycle (Source: Gartner)

Personally, I usually take any of the Gartner data with lots of salt and many a times also have ended up pondering upon their validity. However in this case I would seriously like to believe (or like to pray) that they are right especially in their assumptions about Cloud Computing. Because if they are – I am betting that the real work with proper practical expectations using Cloud Computing would start now. This is the time when many of its potential users would try to cipher through the confusing hype; take baby and practical steps in its adaptation; and let the top spin to see how its application spans out for their needs.

Going forward beyond the peak of the hype, as I am looking into my crystal ball – I am predicting that the usage/adaption of Cloud Computing across products or IT infrastructure would go through the following five stages (see figure below). My assumption of these five stages are derived from what I have typically seen how many other technologies/business processes (e.g. IP Telephony, Enterprise Automation, etc.) have found their way to mass/mature adaption. Note that in the figure below, the curved blue line in the backdrop indicates the Gartner Hype Curve.

Beyond Hype – Predictions about Cloud Computing Adaptations

Beyond Hype – Predictions about Cloud Computing Adaptations

To describe these stages in more details -

  1. Adaption Phase - This phase (and we are well into this phase from Cloud Computing perspective as we speak) is where organizations would start taking the initial steps in looking at the application of Cloud Computing paradigm within their products or infrastructure. Products/applications hosted on traditional infrastructure are getting moved on Public Cloud. SaaS-enablement of existing products is another example. Using Virtualization to create in-house Cloud environment is another activity which many organizations are seriously looking at or many have also started implementing.
  2. Stabilization Phase - As with the adaption of any new technology, the teething problems would need to be overcomed. This is the phase where the myths/hypes/over-expectations associated with the technology would meet with the real-world realities. Various risks associated with the new adaptation would start cropping up and would need to start getting mitigated. To do this certain applications would have to be re-architected; existing business processes would need to be remodeled to find a win-win situations with the business needs and advantages which the Cloud Computing platforms would possibly bring in. I personally believe that this phase would be the most critical of all. Adaptations of many new technologies have struggled in the past in this phase for a long periods and sometimes resulting in the ball getting completely dropped altogether too. From Cloud Computing perspective, I am predicting that the concept itself is going to evolve a lot as organizations go through this phase.
  3. Measurement Phase - Once the adaptation and implementation of any technology stabilizes within any product or organization – the natural next step is to start measuring the day-to-day health and throughput. This is the stage where the Service Level Agreements (SLA) definitions and its measurements is going to start becoming important (as compared to usage of SLAs in the previous stages). At this stage, the measurements associated with uptime, disaster recovery, scalability, response time, etc. would start giving an idea about how well the implementation of Cloud Computing has gone.
  4. Optimization Phase - Measurements would naturally lead to optimization of the implementation – be it from an architecture perspective or simply from the basic notion of how Cloud Computing should continued to be used. It would be too presumptuous of me (or anyone else for that matter) to start predicting how the optimization would play out for Cloud Computing today. It is just too difficult to predict today. However, like all technologies, cloud computing would evolve heavily (for better) in this particular phase.
  5. Management Phase – This is the final phase where success stories associated with successful Cloud Computing implementations would start becoming a regular norm. At and beyond this stage, the implementations of Cloud Computing would continue to focus on regular Management which would include activities such as coordination, reducing regular expenditure, productivity improvement, improving functional statuses and satisfaction.

Please note that I am not suggesting that the above phases are always going to sequential; however it is more from lines of trying to figure out the order of the stages of implementation or even the maturity of the implementation. For example, it certainly does not make any sense to start measuring SLAs even before the implementation has stabilized. In void of such staged approach, I have seen that many a times many false results can dampen the potential which a new technology or paradigm can possibly bring in. And as I have said it before, Cloud Computing is certainly one such powerful platform. However, as many of you would agree with me that even a powerful platform can render itself ineffective if not implemented in a proper manner.

Would love to hear your thoughts/comments.

Microsoft Azure – Its got the muscle but does it have the heart?

July 15, 2009

Within a gap of few months, Microsoft continued its tradition of making a late entry to the party and saying “Hey, I am here too!“. About a month and half back, Microsoft made the announcement of the launch of their new Internet search engine – Bing. Although very late to the party, it seems like it has got some good initial favorable write-ups and reviews (login required). Yesterday they made another such arrival – this time in the Cloud Computing arena – by announcing details about the pricing and details of their upcoming Cloud platform – Microsoft Azure.

Stacey Higginbotham from Gigaom has published a nicely summarized blog – “Everything You Need to Know About Microsoft Azure” – to cover the details of the announcement and what people are saying about it. You may want to refer to it to get the details. Few things that stood out for me in this announcement -

  • Microsoft’s support for allowing the developers to build their products beyond just .NET. Developers can use Azure and build products in Java, Ruby on Rails, Python, etc. My guess is that this is their “Infrastructure as a Service” offering. However, in the “Platform as a Service” offering, .NET Services seems to be their standard and only offering. Still a big step if we were to consider it from Microsoft’s perspective.
  • If I am not completely wrong – they seem to be the first who have come up with a Service Level Agreement (3-9s on the Storage Side and 4-9s on the Compute Side). Non-availability of the SLA had been a common complain against Cloud Vendors like Amazon.
  • There still seems to be substantial reluctance on part of Microsoft to facilitate Azure for implementing Enterprise-wide Private Clouds – however they are saying that they will soon roll-out functionality which will allow users to create a Private-Cloud amongst themselves. In my experience, I have seen Enterprises still very reluctant to move on to the Public Cloud. Although Microsoft certainly has a substantial leverage over enterprises as compared to their peers in the Cloud Computing arena, it would be interesting to see how they succeed in purusading them.

On the whole, on the initial onset there seems to have been a mixed bag of initial feedback with this Microsoft announcement – more seem to be on the positive and hopeful side though. Considering that Cloud Computing itself is at its infancy, IMO it does not matter if Microsoft is relatively late in making this announcement.

Bigger question for me is whether Microsoft has its heart in it from a long term perspective. It is becoming more and more clearer that (and many within Microsoft also have in some admitted to this) that Azure in some way is going to eat into the Microsoft’s primary cash cow – the traditionally packaged software sales. Microsoft has already announced that they would be allowing enterprise customers to take their existing Microsoft licenses into the cloud (in the similar fashion as IBM is also doing for its products on Amazon Cloud). Although Microsoft is betting upon that this entry in Cloud Computing would generate additional revenues for them – what happens if it does not? What happens if additional business does not come but it still hurts the existing selling-licenses business. Would Microsoft still have their heart in Azure in that situation?

That should be the million-dollar question! (I am sure we will see lots of debate on this point for months to come)

I look forward to your thoughts and comments.

Cloud Computing – Is this the case study for Marketing gone out of hand?

June 24, 2009

Over the past few weeks, I have been thinking and doing a lot of reading around what is going around in the Cloud Computing world. About a year back in my group we had started taking some baby steps in getting introduced to this concept and utilizing that to the possible extent in our product development activities. It just seems to me that over this one year organizations are turning over their head to grab a piece of this action. In my own surroundings, I can see many organizations who about a year back had been agnostic to Cloud – suddenly now are in the market claiming that their products are “cloud-enabled“.

In my personal opinion the whole industry is going crazy and the marketing guys are working 24 x 7 to see what kind of cream they can whip out of this milk. McKinsey and Company’s recent report claimed that Cloud Computing is at the peak of its hype cycle and even compared it with the .COM bubble in the early 2000. With the number of definitions of Cloud Computing which are floating around – it just provides an incentive for yet another Industry Expert to create and push forward another new one. The Public Cloud Vendors are cajoling organizations to open their doors and move their applications with them – whereas the Private Cloud Application Vendors are broadcasting that there is no need for that. Any new report/findings which comes out is getting slammed by those whose view points/interests do not match with it. Clearly an indication for the industry to take a deep breather. However, needless to say that my blog post is going to make an iota of difference in that.

In my opinion – Cloud Computing ecosystem should be classified in three discreet buckets ONLY for sake of maintaining simplicity and sanity as listed and defined below (note that I am trying to define these terms very loosely just for basic understanding purposes) -

  1. Infrastructure (Computing and Storage) = Utlilities which provide scalable and on-demand infrastructure required for Computing and Storage and provides it to the consumers in the form on non-intervening services. (Example – Amazon EC2, Amazon S3
  2. Application Platforms = Tools and Technologies together as a platform which enables seamless development (e.g. Google Apps, SalesForce, etc.) or management of applications that run on Cloud (e.g. Appistry, 3Tera, etc.) or use services that is provided from the Cloud (as mentioned in point 1 above) or both.
  3. Software as a Service = Software which are basically constructed and provided more from “usage” perspective. Some examples in this category are services like Gmail, SalesForce.com, etc. (Personally I confess that I am still not completely convinced about this category as I have seen many hosted applications simply claiming themselves as Cloud-enabled. IMO, this is the big ‘elephant’ in the room.)

I think the madness and confusion in the Cloud Computing world has started ever since product marketing has started blurring these three classifications. The attraction and urge for trying to say “me too” seems to have started feeding into this and have created a jungle of definitions and explanations. In my personal opinion, Cloud Computing does not have to be so complex for end-users to find their way around. However, this craze amongst various companies to jump on the bandwagon of “me too” yet “I am different” is making this simple and powerful topic hair-splitting.

In my discussions with my colleagues and friends – many a times the discussions associated with the unrealistic expectations built around Cloud have also come up. In my opinion (and as the Theory of Hype Cycle also says), Cloud Computing may not be able to escape that. However, the bigger concern in my mind is the mis-information and the lack of debate around the awareness of potential challenges (and the possible remedies) associated with Cloud. As an example – the lack of high speed, reliable, and distributed bandwidth can turn out to be a serious impediment in the potential adaption of Cloud.

Cloud Computing is a unique and powerful concept. It has helped in unleashing huge range of product and services and even has the potential of transforming the possibilities associated with existing products. In some sense, we just may have scratched the surface of its potential. However, I do hope that in the rush for getting a piece of the cake – industry also keeps a simplistic view on it.

[Update - October 1, 2009] – Three months after I wrote this blog, I am now getting a feel that the hype around Cloud Computing might be on its decline and now I am thinking that some real work using Cloud Computing may have started. Read more about it in my new blog post titled -”Beyond the hype – can some real work using Cloud Computing start now?


Follow

Get every new post delivered to your Inbox.