Business Daily from THE HINDU group of publications Monday, Nov 23, 2009 ePaper | Mobile/PDA Version | Audio | Blogs |
|
|
|
|
|
eWorld
-
Interview Real promise despite `cloud washing' D. Murali
"... it is the `auto-scaling' aspect that makes cloud very powerful. Auto-scaling is the ability to scale for user volumes or loads linearly."
Suresh Sambandam In Act II, scene III of King Richard III, you hear a citizen say, "When clouds appear, wise men put on their cloaks; when great leaves fall, the winter is at hand; when the sun sets, who doth not look for night?" An apt quote, I feel, looking back at my meeting with Suresh Sambandam, Founder & CEO, OrangeScape Technologies Ltd, Chennai (www.orangescape.com), a few weeks ago in Nageswara Rao park. For, `clouds' did appear in our discussion that stretched right through the dusk, though whining insects were at hand, rather than the winter. There is a lot of hype going around about `Cloud Computing' and it seems to mean different things to different people, Sambandam begins. "Cloud is the new fashion word in tech world and many vendors are `Cloud Washing' their products even though they have nothing to do with cloud." What is cloud computing, I ask him? At an abstract level, it is about two key things, explains Sambandam. "The ability to: 1) process information - think CPU; and 2) store information - think disk." The computing needs are ever-growing whether it is to render high-end games for personal entertainment or heavy-duty analytics for enterprise MIS, he continues. "The PCs at home and servers at work are not able to scale up to this demand. Most times, the need for such high-end computing is spiky and the ROI for investing in high-end computing just for the spikes doesn't make sense." So? What if you can hire computing resources (CPU/disk) on need basis and pay for the actual usage, asks Sambandam rhetorically? Just like the way we are using electricity at home, I wonder? That is cloud computing for you, he responds. "The basic premise is that, infinite amount of computing resources (CPU/disk) are hosted somewhere in the cloud (Internet) and you can access them on-demand and pay for what you actually use." Our conversation continues over e-mail. Excerpts from the interview: What are the benefits of cloud? There are two key benefits that a cloud provider should offer: 1) On-demand capacity; and 2) auto-scaling. On-demand capacity is the commonly talked-about aspect of cloud. However, in my opinion, it is the `auto-scaling' aspect that makes cloud very powerful. Auto-scaling is the ability to scale for user volumes or loads linearly. More simply put, there should not be any noticeable degradation in performance or response times of your software whether it is being used by 10 users or 10,000 users concurrently. Also this question naturally leads us to understand the different layers in cloud, viz. IaaS (Infrastructure as Service), PaaS (Platform as a Service), and SaaS (Software as a Service). SaaS and cloud are used interchangeably. True, SaaS and cloud are often interchangeably used. But, cloud is a delivery platform and SaaS is a business model. SaaS is about providing software as a service instead of licensing the software. Also, cloud makes SaaS feasible and hence cloud and SaaS kind of go hand in hand in most cases, but there are a few exceptions. SaaS is a great boon for consumers and enterprises. In the enterprise software space, the vendor delivers the CD to the customer and the fun begins for the customer from there. The customers have to spend months together in turning a CD into a software that goes live into production. In contrast, SaaS puts all the ownership on the software provider to deliver software that is working, usable and constantly relevant, i.e. upgraded continuously. Consumers and enterprises start realising the value for their money from the word `Go'. Also, I strongly believe that SaaS is the `industrialisation' equivalent for the software industry. We keep hearing about multi-tenancy in the SaaS space. What is it? Most software companies, who have existing applications, especially in the business to business (B2B) space, think of SaaS-ifying their application by hosting on their own hardware or on an IaaS platform such as AWS or Rackspace. This is not true SaaS. This is plain hosting or ASP equivalent that happened during the dotcom days. One of the key characteristics of a B2B SaaS application is about multi-tenancy, which is the ability to run the same software for multiple customer datasets on the same hardware, thereby amortising the hardware costs across multiple customers. Only those companies that do this correctly can see light at the end of their profitability tunnel. Sadly, there is no magic bullet for SaaSification. The existing application has to be re-architected to support multi-tenancy to become a `true SaaS' provider. Multi-tenancy is not so much of a problem if you are a business to consumer (B2C) software provider. Is cloud, like dotcom, another bubble waiting to burst? Watching what is happening closely in the SaaS/cloud space will tell you that this is real and no hype. Cloud/SaaS is a tectonic shift for IT industry as a whole. For example, Siemens moved 4,50,000 of its employees worldwide into a SaaS-based software for performance management. This is just one example of large-scale adoption of SaaS. SalesForce.com, a single company that operates in the SaaS model, makes close to $1 billion every year; if you put all other companies that operate in the cloud/SaaS space, the numbers are quite big. These numbers suggest that large-scale adoption is happening, leading to revenues for vendors. This did not happen during the dotcom bubble. Real money is being spent, which makes it a huge difference. What is the impact of cloud/SaaS on chief information officers (CIOs) of large enterprises? Cloud is going to make hardware even more a commodity. Large enterprise CIOs, i.e. the likes of Unilever, Pfizer etc., will go for their own private clouds whereas mid-market CIOs will go in for virtual private clouds. These two paradigms (SaaS/Cloud) are going to help forward-thinking CIOs to implement IT initiatives that are in line with the timelines of the business. These bring in agility with lower TCO (total cost of ownership). These are long-pending unfulfilled demands of business from IT. CIOs would choose SaaS offerings for undifferentiated applications and go for bespoke development only for applications that are unique and are a competitive differentiator. The cost of automating business process into IT systems will go down with cloud/SaaS and hence whatever that didn't make ROI sense earlier will make sense now. Early mover CIOs will have a great advantage. What is the significance of SaaS as regards small businesses? The cost of good quality software was always on the rise until we hit SaaS. Enterprise software, being a heavy capital purchase, makes it a premium play for small business and hence they weren't able to afford it. SaaS is a great equaliser. Good quality enterprise software is now affordable at monthly operating costs for SMBs. This enables SMBs to compete equally with their industry giants on their core value propositions Do you see cloud/SaaS creating new business opportunities? Definitely. SaaS is a boon, particularly for the Indian IT industry. Conventional B2B vendors take a software + service approach. For example, 60-70 per cent of the functionality is provided by the ERP platform and the remaining needs to be customised using services. Typically, the cost of licence to services ratio is in the order of 1:2.5. This clearly means that without significant spend on services, the value from the B2B enterprise software (ERP) can't be derived. The reason of this large amount of services is that these software are developed generically to address the needs of all geographies and verticals. The processes of a real-estate company are quite different from that of a manufacturing company. Also, the real-estate companies in Dubai would operate differently when compared with those in India. So, generic stuff doesn't work. On the other hand, SaaS is about a focused application for a specific industry vertical and for a specific geography. The 3-Dimensional interaction for an application, industry and geography opens up an opportunity for a potential SaaS application. These are up for grabs for the Indian IT services companies as we have the manpower which no other country has. If we can ride this wave, we can change the narration of India from a `Low Cost IT Services Destination' to becoming `IP Lead' through SaaS products. What is the impact of cloud on Government services such as e-Gov and NeGP? Government is the single-largest IT buyer. Central Government alone purchased $3 billion worth of IT hardware, software and services last year, and this is expected to go to $5.1 billion in 2011. In my opinion, Government can save a lot of money by using cloud. Roughly 50 per cent of Government purchases are in hardware and mostly they are under-utilised. The cloud gives pay per use model there by avoiding unnecessary capital investment. Also, a large number of hardware investments by Government suffer from obsoleteness. Government should consider adopting the US model of putting out a common cloud for Federal Services and specific clouds for State-level services. The existing SDC (State Data Centre) investments part of the NeGP (National e-Governance Programme) should really be used towards this purpose. We hear of concerns about data security and privacy of data in Cloud. What is happening on this front to achieve customer confidence? The real truth is data is very insecure with enterprises and individuals. What if you lost your laptop, and your backup is one month old? You are doomed. Today, I put most of my data in the cloud, except for those applications that are not cloud ready yet. I feel safe and stress free. We are going through a phase of change where we are asked to put our data in someone else's computer. It is the resistance to this change that is creating all these questions in people's mind. In olden days, we stored valuables underground in our own houses or some secret place. Today, we put our valuables like gold in our banker's locker. This is because we have realised that we can't invest the money to build all the safety and security systems required to replicate the bank locker at our homes. In fact, in many cases, the cost of such a secure locker will be more than the money we have. This is no different for data. Except for very large enterprises nobody can afford to build a secure and safe data centre. Using a SaaS application is like using a compartment in your banker's locker. Currently, we are going through the awareness phase and over time this will surely settle down. It is only a matter of time. Does cloud mean anything to the common man? I understand that there are 495 million mobile subscribers in India when compared with just 5 million broadband connections at homes. This clearly means that Internet adoption is very poor. The reason is that there is no killer application for Internet adoption to happen. In my opinion, SaaS is a killer application for Internet, especially for India. SaaS will drive Internet usage like the way we use mobile phones. You'd mentioned IaaS and PaaS. IaaS provides necessary hardware infrastructure to run your software. This space is a heavy investment game. Amazon's AWS offering is the leader in this space accompanied by RackSpace, GoGrid and other data centre players such as NetMagic in India, who are also quickly jumping into the IaaS play. From a technical perspective, this works on a virtualisation approach, i.e. taking a huge machine and allocating small virtual machines for you on request. There are some inherent challenges in achieving auto-scaling with this approach, as scaling can't be achieved without understanding your application characteristics. However, using this approach existing Web applications that are already written on Java/J2EE or .Net or open source languages such as Python, PHP or Ruby can be easily moved on this to satisfy increasing capacity needs. Simply put, this is nothing other than a hardware play but on the `Cloud'. In the PaaS (Platform as a Service) space, there are two categories of players, i.e. the companies that provide `Cloud OS' and the companies that provide application platform / cloud middleware. Google with its App Engine (GAE) product operates in the `Cloud OS' space. Google's offering is more like an operating system that unifies a large number of smaller computers together, more commonly known as commodity hardware (or PCs), creating a big mammoth computer. This is a grid model, the opposite of Amazon's virtualisation model. There are strong restrictions with respect to the choice of the programming language. Python and Java are the only supported languages. Also no other option, other than using Google's proprietary database called `BigTable'. While GFS/BigTable is proprietary it is quite powerful and quite easily is the largest data store in the world. Even though GAE has limitations, it clearly meets both the characteristics of a `Cloud Provider' i.e. on-demand capacity and auto-scaling, thumbs up when compared with any other provider. Even if your existing software is written on Java o Python, you have to rewrite good portions of your existing software to run on GAE platform. It would be a worthwhile exercise, if you are anticipating tens of thousands of users hitting your system. Microsoft's yet-to-be-released Azure platform also maps on to this space and would be a competition to GAE. Another interesting space to watch is the application platform or cloud middleware space. These products typically contain a full range of capabilities, such as rules engine, UI Engine, data storage engine, workflow engine and integration engine all packaged into one pre-integrated product offering. These products often provide dramatic time reduction in learning curve, time to market and effort required to make ongoing changes. While these products are less flexible when compared with a free form programming language like Java, it is worth a trade off. These products will be an ideal choice for entrepreneurs or business savvy individuals who want to turn their expertise into a SaaS based software. In fact, these products have a great potential to spur up the SaaS start-ups in India creating a mushroom of what I call micro-ISVs. In addition to that, large enterprises can use these platforms for building large scale extended enterprise applications for servicing their suppliers and customers. More Stories on : Interview | Software
Article E-Mail :: Comment :: Syndication :: Printer Friendly Page
|
|
The Hindu Group: Home | About Us | Copyright | Archives | Contacts | Subscription Group Sites: The Hindu | The Hindu ePaper | Business Line | Business Line ePaper | Sportstar | Frontline | The Hindu eBooks | The Hindu Images | Home |
Copyright © 2009, The
Hindu Business Line. Republication or redissemination of the contents of
this screen are expressly prohibited without the written consent of
The Hindu Business Line
|