Ten things I learned while writing Cloud Computing for Dummies

I haven’t written a blog post in quite a while. Yes, I feel bad about that but I think I have a good excuse. I have been hard at work (along with my colleagues Marcia Kaufman, Robin Bloor, and Fern Halper) on Cloud Computing for Dummies. I will admit that we underestimated the effort. We thought that since we had already written Service Oriented Architectures for Dummies — twice; and Service Management for Dummies that Cloud Computing would be relatively easy. It wasn’t. Over the past six months we have learned a lot about the cloud and where it is headed. I thought that rather than try to rewrite the entire book right here I would give you a sense of some of the important things that I have learned. I will hold myself to 10 so that I don’t go overboard!

1. The cloud is both old and new at the same time. It is build on the knowledge and experience of timesharing, Internet services, Application Service Providers, hosting, and managed services. So, it is an evolution, not a revolution.

2. There are lots of shades of gray with cloud segmentation. Yes, there are three buckets that we put clouds into: infrastructure as a service, platform as a service, and software as a service. Now, that’s nice and simple. However, it isn’t because all of these areas are starting to blurr into each other. And, it is even more complicated because there is also business process as a service. This is not a distinct market unto itself – rather it is an important component in the cloud in general.

3. Market leadership is in flux. Six months ago the market place for cloud was fairly easy to figure out. There were companies like Amazon and Google and an assortment of other pure play companies. That landscape is shifting as we speak. The big guns like IBM, HP, EMC, VMware, Microsoft, and others are running in. They would like to control the cloud. It is indeed a market where big players will have a strategic advantage.

4. The cloud is an economic and business model. Business management wants the data center to be easily scalable and predictable and affordable. As it becomes clear that IT is the business, the industrialization of the data center follows. The economics of the cloud are complicated because so many factors are important: the cost of power; the cost of space; the existing resources — hardware, software, and personnel (and the status of utilization). Determining the most economical approach is harder than it might appear.

5. The private cloud is real.  For a while there was a raging debate: is there such a thing as a private cloud? It has become clear to me that there is indeed a private cloud. A private cloud is the transformation of the data center into a modular, service oriented environment that makes the process of enabling users to safely procure infrastructure, platform and software services in a self-service manner.  This may not be a replacement for an entire data center – a private cloud might be a portion of the data center dedicated to certain business units or certain tasks.

6. The hybrid cloud is the future. The future of the cloud is a combination of private, traditional data centers, hosting, and public clouds. Of course, there will be companies that will only use public cloud services for everything but the majority of companies will have a combination of cloud services.

7. Managing the cloud is complicated. This is not just a problem for the vendors providing cloud services. Any company using cloud services needs to be able to monitor service levels across the services they use. This will only get more complicated over time.

8. Security is king in the cloud. Many of the customers we talked to are scared about the security implications of putting their valuable data into a public cloud. Is it safe? Will my data cross country boarders? How strong is the vendor? What if it goes out of business? This issue is causing many customers to either only consider a private cloud or to hold back. The vendors who succeed in the cloud will have to have a strong brand that customers will trust. Security will always be a concern but it will be addressed by smart vendors.

9. Interoperability between clouds is the next frontier. In these early days customers tend to buy one service at a time for a single purpose — Salesforce.com for CRM, some compute services from Amazon, etc. However, over time, customers will want to have more interoperability across these platforms. They will want to be able to move their data and their code from one enviornment to another. There is some forward movement in this area but it is early. There are few standards for the cloud and little agreement.

10. The cloud in a box. There is a lot of packaging going on out there and it comes in two forms. Companies are creating appliance based environments for managing virtual images. Other vendors (especially the big ones like HP and IBM) are packaging their cloud offerings with their hardware for companies that want Private clouds.

I have only scratched the surface of this emerging market. What makes it so interesting and so important is that it actually is the coalescing of computing. It incorporates everything from hardware, management software, service orientation, security, software development, information management,  the Internet, service managment, interoperability, and probably a dozen other components that I haven’t mentioned. It is truly the way we will achieve the industrialization of software.

The end of maintenance?

I admit that I didn’t read the whole article but then I really didn’t have to. I knew what Marc Benioff, CEO of Salesforce.com was trying to start. I remember many years ago seeing Marc at an industry conference where he proudly announced the end of software.  A nice marketing approach that definitely got everyone’s attention. Of course, at that time Marc was working on a little software as a service enviornment that became Salesforce.com. The rest is history, as we like to say.  Now, Marc is on a new mission to attack maintenance fees. While it is clear that Marc is trying to tweak the traditional software market I think that he is bringing up an interesting subject.

Software maintenance is not a simple topic to cover and I am sure that I could spend hundreds of pages discussing the topic because there are so many angles. Maintenance fees began as a way of ensuring that software companies had the revenue to fund development of new functionality in their software products. It is, of course, possible to buy software, pay once, and never pay the vendor anything else. Those situations exist of course. Ironically, the better designed the software, the less likely it is that customers will need upgrades. But, clearly that circumstance is rare.

There are major changes taking place in the economics of software. Customers are increasingly unhappy with paying huge yearly maintenance fees to software providers. Some of these fees are clearly justified. Software is complex and vendors are often required to continue to upgrade, add new features, and the like. There are other situations where customers are perfectly happy with software as is and only want to fix critical problems and don’t want to pay what they see as exorbitant maintenance fees.

Now, getting back to Marc Benioff’s comments about the end of maintenance. Here is a link from Vinnie Mirchandani’s recent blog on the topic.Marc is making a very important observation. As the world slowly moves to cloud computing for economic reasons there will be a major impact on how companies pay for software. Salesforce.com has indeed proven that companies are willing to trust their sales and customer data to a Software as a Service vendor. These customers are also willing to pay per user or per company yearly fees to rent software. Does this mean that they are no longer paying maintance fees? My answer would be no. It is all about accounting and economics. Clearly, Salesforce.com spends a lot of money adding functionality to its application and someone pays for that. So, what part of that monthly or yearly per user fee is allocated to maintaining the application? Who knows? And I am sure that it is not one of those statistics that Salesforce.com or any other Software as a Service or any Platform as a Service vendor is going to publish. Why? Because these companies don’t think of themselves as traditional software companies. They don’t expect that anyone will ever own a copy of their code.

The bottom line is that software will never be good enough to never need maintenance. Software vendors — whether they sell perpetual licenses or Software as a Service– will continue to charge for maintance. The reality is that the concrete idea of the maintenance fee will evolve over time. Customers will pay it but they probably won’t see it on their bills.  Nevertheless, the impact on traditional software companies will be dramatic over time and a lot of these companies will have to rethink their strategies. Many software companies have become increasingly dependent on maintenance revenue to keep revenue growing.  I think that Marc Benioff has started a conversation that will spark a debate that could have wide ranging implications for the future of not only maintenance but of what we think of as software.

Does IT see the writing on the cloud wall?

For the last six months or so I have been researching cloud computing. More recently, our team has started writing our next Dummies Book on Cloud Computing. Typically when we start a book we talk to everyone in the ecosystem — vendors big and small and lots of customers.  For example, when we started working on SOA for Dummies almost three years ago we found a lot of customers who could talk about their early experience. Not all of these companies had done things right. They had made lots of mistakes and started over. Many of them didn’t necessarily want their mistakes put into a book but they were willing to talk and share.  As I have mentioned in earlier writings, when we wrote the second edition of SOA for Dummies we had a huge number of customers that we could talk to. A lot of them have made tremendous progress in transforming not just their IT organization but the business as well.

We had a similar experience with Service Management for Dummies which comes out in June. Customers were eager to explain what they had learned about managing their increasingly complex computing and business infrastructures.  But something interesting in happening with the Cloud book. The experience feels very different and I think this is significant.

Our team has been talking to a lot of the vendors — big and small about their products and strategies around the cloud. Some of these vendors focused on some really important problems. Others are simply tacking the word cloud in front of their offerings hoping to get swept up in the excitment. But there is something missing. I think there are two things: there is a lack of clarity about what a cloud really is and what the component parts are. Is it simply Software as a Service? Is it an outsourced infrastructure? Is it storage capacity to supplement existing data centers? Is it a management platform that supports Software as a service? Does cloud require a massive ecosystem of partners? Is it a data center with APIs? Now, I am not going to answer these questions now (I’ll leave some of these to future writings).

What I wanted to talk about was what I see happening with customers.  I see customers being both confused and very wary. In fact, the other day I tried to set up a call with a senior executive from a large financial services company that I have spoken to about other emerging areas. This company always likes to be on the forefront of important technology trends. To my surprise, the executive was not willing to talk about clouds at all.  Other customers are putting their toes in the cloud (pun intended) by using some extra compute cycles from Amazon or by using Software as a Service offerings like SalesForce.com. Some customers are looking to implement a cloud-like capability within their own data center. Could it be there they are afraid that if they don’t offer something like Amazon’s EC2 cloud that they will be put out of business? Just as likely they are worried about the security of their intellectual property and their data.

I predict that the data center is about to go through a radical transformation that will forever change the landscape of corporate computing. Companies have recognized for a long time that data centers are very inefficient. They have tried clustering servers and virtualizing their servers with some level of success.  But the reality is that in time there will be a systematic approach to scalable computing based on the cloud.  It will not be a simple outsourced data center because of the transition to a new generation of software that is component based and service oriented. There is a new generation of service management technologies that makes the management of highly distributed environments much more seamless. The combination of service oriententation, service managment, and cloud will be the future of computing.

The bottom line is that while the vendor community sees dollar signs in this emerging cloud based world, the customers are afraid. The data center management team does not understand what this will mean for their future. If everything is tucked away in a cloud what is my job? Will we still have a data center? I suspect that it will not be that simple. At some point down the line we will actually move to utility computing where computing assets will all be based on a consistent set of standards so that customers will be able to mix and match the services they need in real time. We clearly are not there yet. Today there will be many data center activities that either cannot or will not be put into a cloud. Internal politics will keep this trend towards clouds moving slowly.

What happens to SaaS in a tough economy?

I participated in a SaaS event this week that was sponsored by IBM.  It was sort of a funny feeling to be at a very good, positive event that focused on SaaS as a platform right in the midst of an economic meltdown.  In some ways, I had one of those out of body experiences. What am I doing talking about the future of SaaS when the world seems to be crashing and burning. As we sat listening to speakers and talking to each other the stock market went down 700 points. I met many different software executives from companies that are creating very significant SaaS based offerings — and they are getting good traction from their customers.

But the question remains and one that I will attempt to answer is what will happen to SaaS in this economy. I think that SaaS is going to be hugely successful in this economy. First, it is clear that customers are growing increasingly comfortable with the idea of using software that is managed by a third party vendor and hosting provider.

Now not all these vendors are equal. It is actually tricky to ensure success in a SaaS world.  After all, if you buy a regular software license and then decide that the software is not as good as you thought, you are stuck. Now, next year you might forgo the maintanance fee, but you still own the code.  It is different with SaaS. If you decide to take on the 30 day free trial there is no guarantee that you will become a life long customer. Likewise, if you do take the plunge and sign on for a month or two, there is also little guarantee that you will become devoted to the application. My point is that becoming a good, profitable and predicatable SaaS vendor is harder than it looks. Basically, you’ve got to be pretty good to make it.

Now, back to the economy and SaaS. Customers who will still need software even in a horrible market are going to think twice about captial expenditures.  Do you really want to spend a lot on servers and storage and the like? I predict that in tough economic times paying someone a monthly or even a yearly fee and letting them buy the capital intensive stuff will be just the ticket.

So, I think you will see the really smart SaaS vendors that know how to proactively nuture their customers so that they will really use their technology will win.  These smart SaaS vendors will also figure out the meaning of scalability, performance, and managability.  They are already figuring out how to make their software configurable and they are even creating versions that appeal with vertical market segments.

This economic climate may be making us all a little crazy and scared but there are some nice opportunities for those who are willing to solve customer problems. This will be the beginnng of the SaaS renissance and I think it will be a positive move for customers and the market.

An evening with Salesforce.com

I was invited to a dinner with some of the management team last night. This was the third year in a row that the company had held a little get together for key customers, IT analysts, and members of the press. In previous years, CEO, Marc Benioff himself to lead the festivities. This year, he was called away and wasn’t able to join us…oh, well. I will have to wait until next year. I actually first met Mark when he was a business unit head at Oracle. He left Oracle in the late 1990s to start Salesforce.com. In fact, I was one of the early testers of the beta version of Salesforce.com. It was certainly a different product in those days when it was focused on the single end user who wanted to keep track of contacts in a usable way.

The company has come a long way. It is impressive to look at what they have accomplished in a few short years. Salesforce.com not only has 35,000 customers but it has created an ecosystem of partners within the realm of software as a service. In fact, I am seeing established enterprise software companies such as Informatica becoming partners (for data cleansing). Even more interesting is the raft of startups that are tying their fortunes to the salesforce.com platform.

At the dinner meeting I had the opportunity to sit next to the CIO of an important Pharmaceutical company who informed me that software as a service is becoming the lynch pin of the company’s software strategy for non-security related tasks. The company heavily uses both Salesforce.com and Workday as its ERP system. The main reason for choosing SaaS: the need to reduce expenses both for licenses and personnel.

In my view the question is not whether Software as a Service is real — it is. The question is how well the new players like SAP, IBM and Microsoft will perform as they point their marketing machines at the market.

SOA and Unintended Market Consequences

Now, I have been around the software world for longer than I would like to admit (I would be giving away my age if I told you). I have watched many trends come and go. The one consistent pattern that I see is the slow progression over the last 20 years towards distributed computing. There has been a lot of innovation – some that has resulted in some pretty interesting products and companies that made their mark on the world and then went away.

 

The advent of service-oriented architectures is not simply another fad with another set of products and services. I really believe that it will have the same long-term impact as the Internet did when it arrived on the scene as a commercial venture in the early 90s. If one thinks back to those days, many people had no idea of how this quaint piece of technology would actually be used in the commercial world. And, as we say, the rest is history.

 

Just as the Internet produced unintended consequences for the future of computing, I expect that SOA will too. It is my prediction that Service-Oriented Architectures will force a dramatic change in the balance of power in the software world – both on the vendor and the customer side. Why do I make this rash statement? Start by thinking about the implication of SOA. It is a dramatic change in the way software is created, reused, recombined, managed, and sold. From a customer perspective, SOA puts power in the hands of the business user.

 

So, what could change? First, in time, SOA will level the playing field. Today the major infrastructure vendors are all vying for control over who owns the customer’s infrastructure. Will SAP, Oracle, HP, IBM, Microsoft or an open source play, like JBoss be the winner? Will one of these companies convince enough customers to do it their way? Perhaps.

 

But I am envisioning a different possible scenario. Imagine that we get to the stage where vendors converge around codified standards such as XML and everything necessary for creating this highly distributed architectural framework. Imagine that there are enough good adapters that allow components from various environments to easily connect to each other. And now imagine that what customers really want to do is to reuse their existing software assets by encapsulating them into business services that are used over and over again to create new virtual applications.

 

If you follow my thinking here, it opens up interesting opportunities. Here are a few thoughts worth considering:

 

• Customers will take SOA seriously – not just as a technology initiative but as a business strategy – they will demand that industry leaders cooperate to make the customers’ job easier.
• New vendors will emerge that can leverage the SOA work done by infrastructure leaders to provide commercialized services address specific problems in specific vertical markets.
• The first era of the industrialization of software is initiated. Vendors will begin to create modular offerings based on a SOA model that are designed to fit into existing frameworks (kind of like steering wheels that can be used on 20 different car models).
• The packaged application that has been the mainstay of the software industry for the past 25 years begins to transform into these industrial packaging. In time – give us another 10 years – the traditional software package disappears.
• SOA opens the door to a new world where Software as a Service is the norm. This transition will take at least 15-20 years. But it is real and will have a dramatic impact on the entire industry from financial models to the importance of comprehensive hosting providers.

 

This is a lot to consider. A lot of companies are trying to figure out whether SOA is real or just a passing fad. Will the words Service-Oriented Architectures soar as a lynch pin of the future or be edged out by new hot silver bullets? Given the history of the computer industry, it’s not surprising that people are nervous. After all, trends come and go at a rapid pace. I predict that this one is real. But let me warn you: it is not a quick fix. We are, in fact, at the beginning of a big emerging market trend that will have ripples in the software market for the next decade and beyond.