Infosys-Oracle Blog

Cloud is indeed the future. It gives companies immense benefits in terms of scalability, ease of IT expenditure - conversion of capital expense to operational expense etc? However, it makes good business sense for service providers to offer such standardized service when given to a large number of clients. In such scenario, how do you view the role of system integrators like Infosys in the future scenario of standard applications - with little or no customization - delivered on the cloud?

Hi Ravi, thanks for sharing your thought on Cloud.
A powerful component of Cloud applications is that such a sophisticated interface and accompanying data can easily be accessed through a web browser and an internet connection. The end user does not need to be concerned with the underlying technology and where or how the data and applications are stored.
Cloud and SaaS together provides a very powerful and cost effective platform for our clients. Wanted to know your thoughts on how bigger and established clients are thinking about migrating to cloud.

As per Sep 2009 IDC report, Cloud computing is expected to grow at 24% CAGR to $ 42.3 billion market by 2012. Oracle, SAP, Microsoft, Amazon, SalesForce.com, Google, Lawson are already investing heavily in the Cloud Offerings.

However, this computing model is still in evolution stage and has to overcome challenges (some real, some perceived!!!) like security, interoperability, data mobility, compliance and business continuity etc. before it becomes the model of choice. Also, there is no single Cloud computing model that will fit all of the enterprises requirements and always. Thus, enterprises need to do a comprehensive evaluation before adopting the relevant Cloud model for their needs (IAAS/PAAS/SAAS, Private vs. Public Cloud).

• Large enterprises having large IT spend and mature IT model, will typically not be enticed upfront to join the Cloud bandwagon. However they can periodically check their on-premises’ capacity and plan Cloud Virtualization for additional workloads. Also Private clouds will typically be more conducive for them viz. Public clouds. They will be tempted to move their Non Strategic functions like MDM etc to cloud ahead of their core functions.

• Large and growing enterprises having large IT spend but less mature IT model, are likely to adopt cloud to do a rapid catch up on their IT side so that the IT can support their fast business growth. Also the cost savings by moving to Cloud will allow them to divert their IT budget to building new capabilities and innovation on the business side.

• Small enterprises but having mature IT model, will be looking to curtail IT spending by consolidating the IT application spread. Infrastructure consolidation and virtualization will help bring down cost. Enterprise can buy storage and other IT functions as and when need. This brings in significant flexibility in managing IT investments. Public cloud having multi tenant architecture is a viable option for these scenarios.

• Small enterprises and having less mature IT model, will have their business processes being managed using conventional and manual methods. IT strategy will be to automate business to reduce redundancy and improve efficiency. In this segment, Cloud ERP multi tenant architecture with pay per use model can be a fit.

Though the concept of Cloud computing is simple, the impact and mechanisms for delivery are far more complex. Enterprises need to do a detailed analysis of long term cost and benefits before moving to Cloud. Cloud service providers need to keep innovating and make customer experience more enriching then before.

- Abhishek Goyal, Oracle @Infosys Technologies

Cloud is the future. Infact people are telling except for the core application like data on banking industry which are highly sensitive, everything will be on cloud in the near term. This application makes more sense for emerging economies where business users are price sensitive and cost of capital is pretty high. Infact we will gradually see, increasing convergence of cloud computing ,enterprise mobility solution and social media networking. Cloud will not only give scalability but huge cost and asset optimization depending on needs and requirements of the business users. In fact it is becoming clear that over a period of 3-4 years and beyond, the benefits of cloud computing will override the benefits accrued on TCO model . Also this another model which is game changing in nature and will empower a lot of competitors small ,medium and large to offer unique and point to point solutions.

One of the most important challenges on the critical path ,to bring any enterprise on Cloud is how quickly and efficiently one can integrate the existing on-premise infrastructure of the business with the cloud solutions. Cloud Users have realized there must be some close handholding between on premise and cloud infra to maximize business value. So Cloud Service Providers are needed to brace for a lot of questions on how they can make this hybrid model (on premise and cloud) to work and reshape the core IT processes of the incumbent and prospects

While cloud computing has been prevalent in the consumer & SMB space, these are still early days in the enterprise space. There has never been a compelling business case for enterprises than now:
- Lower upfront costs (Opex Vs Capex), thereby making the business case viable.
- Optimization of data center costs via virtualization. Data centers account 25% of total corporate IT budget.
- Sustainability. Carbon emissions from these data centers contribute to 0.6% of world wide emissions!!! Data center emissions are
expected to quadraple by 2020.
- On-Demand scalability and availability.

Many companies are utilizing cloud services today. The market has progressed from Infrastructure as Service (Iaas), Software as a
Service (SaaS) to Platform as a Service. There are multiple delivery models which have emerged - Enterprises are adopting Private clouds to ensure there is no compromise on data security.

While it is important to have a vision for cloud computing, it is not the end goal for organizations. It is a means to achieve their busienss objectives aka 'Value'.

Great questions and thoughts and let me give my views on each of them:
1. The constraints around security, adherence to slas, customization needs of enterprise and mission critical applications on cloud will slowly be overcome and overwhelmed by the profound advantages the cloud would bring and make technology faster, easier, more flexible and more effecctive. The case is so compelling that we will see faster innovation to overcome these challenges.

2. Economies of scale because of supply side savings and demand aggregation will drive adoption on the cloud. Supply side savings will be driven by reduction in cost of power, infrastructure support and maintenance costs and bargaining power of large data centers. Therefore the cost of capacity will go down. Demand aggregation would be driven by virtualization that enables multiple applications to run in the same environment. Optimizing utilization of resources like CPU, Network and Storage through virtualization would lead to significant savings. Clearly how much you can leverage through higher economies of scale depends on how much your workloads are variable over time. In real life situations we would require large number of resources in a particular time and literally none in a different time window therefore opening up opportunities to diversify/share resources across depts, companies, Industries etc and hence reduce costs through the concept of cloud.

Interesting perspective.
All this shows that Cloud computing is inching towards IT commoditization. This reminds of the famous article that Nicolas Carr has written way back in 2003 on “IT doesn’t matter”.
Cloud is fast becoming a business model, not just technology option. No doubt, it will take time ( depending on the size of the organization infrastructure consolidation and virtualization can take up anywhere between 1 to 3 years). What is also interesting is the myriad of possibilities & players on the next level of cloud based business models – platform, software / application.
Nicolas Carr explained this significance of cloud computing in his book, “The Big Switch”, and elucidated on how companies switched, more than hundred years back, from self-power generation to cheap power generated by newly built electric grid. A close look at what happened in the semi conductor industry, may be a harbinger of things to come – while at one point every hardware vendor had a captive fabrication unit, that is no longer the case now. Due to requirements of scale, efficiency and result of aggregation not many companies ( exception: Intel, Samsung) can afford a captive semi-conductor fabrication facility ; this led to the rise of semi-conductor foundries (like TSMC) which builds chips for others.
Amazon with services like EC2 and Sales-force are the initial movers in the cloud services space. Google and Microsoft are investing in huge funds in data centres. IBM, Dell, HP and Oracle are not far behind and are actively targeting enterprise market. Cloud based business model is also one of the biggest threats for existing players – fundamental disruption of the complete Value chain ( Storage / HW Vendors, ISVs, Service Providers and IT Depts). Will the cloud providers be the “foundries” of the morrow ?

Cloud is a great concept. However there are hurdles that need to be overcome. Two examples-
1. Interoperability: A major challenge from a customer perspective. Can competitors collaborate? What will prevent a lock down from a single vendor?
2. Accountability : BP Oil Spill is a classical example. At the end of the day the customer is responsible for his information. Contracts could mitigate this, but the customer loses some level of control. Related are quality aspects. Who is accountable / responsible for the quality of the information? The customer who enters the information or the vendor who provides the solution. These are grey areas.

On the whole I think that the cloud concept will be influenced by how major players are able to influence the regulatory bodies. We must realize that the challenges are technical and non technical. While smart people can resolve the technical challenges, it is the non technical ones that will influence the way the cloud concept is perceived and accepted by customers.

I like the thoughts in "Big Switch" and would recommend people to read on how adoption on electricity grids as a utility got embraced by corporates. I see the Cloud evolving in the same way.

I wanted to give a slightly detailed perspective on "Economies of Scale" to give a view on the Interoperability part. While supply side economies described in my previous note is a no-brainer, demand aggregation is a very interesting area to probe into. Lets talk about a few categories:
1. Industry variations: Variations on computing needs across Industries can be aggregated and leveraged on the cloud into specific clusters. For example: Retail Industry has seasonal variations which are different to needs of Utilities companies and hence we could aggregate infrastructure and technology needs of customers in these two Industry verticals

2. Time of the day variations: Daily recurring cycles with peaks on specific time slots in a day can lead to unused capacity for the rest of the day. We could therefore aggregate infrastructure needs of a group of customers who have different peaks during the day.
3. SLA needs: To meet SLAs, capacity buffers have to be built into organizations. By pooling requirements across organizations you could reduce the probability of occurance of peak demand of infrastructure resources

Ravi and Friends, nice to see interesting and informative thoughts, shared through this blog/discussion forum on Clod Computing.

Having spent most of my time in the telecom domain, I wish to share the fact, that the starting point of the cloud concept was with the evolution of telecom networks, where you will always see the telecom networks represented as a Cloud. Even in concept of land line or mobile networks is fundamentally a Cloud, where the telephone exchange or switching system is positioned at the service provider premises and networked across the country/globe and the services are offered to the end customers (who just own their telephone devices and not the telecom infrastructure and they just pay for the usage of telephone services).

Similarly, in the Cloud computing concept (SaaS/PaaS/IaaS) are similar, where the computing infrastructure is positioned as a cloud and the IT Applications/Services are offered to the end customers (who just need to own their computers/smart phones for access and they need not own the IT infrastructure and they just need to pay for the usage of the IT applications/services).

As the Cloud concept has been well proven and succeeded in the Telecommunication world, over the last 100+ years, certainly the concept of Cloud Computing will continue to progressively evolve/ mature, succeed and grow phenomenally in the next 100+ years in the IT World.

One other area which would drive adaptability of cloud and leverage its advantages is the evolution of multi-tenancy as a concept for business applications. Innovation and research would drive greater push towards reducing the constraints around multi-tenancy and therefore push more acceptability. Multi-tenancy would therefore drive sharing of system instances and technology infrastructure across customers for key functional areas. To start with the applications associated with back office functions will adopt faster. Ofcourse Larry Elison doesn't agree to this, as his belief is that virtualization is so powerful, evolved, cheap and mature that you don't need to evangelize multi-tenancy

As we discuss demand aggregation as an Important driver for cloud computing, I wanted to discuss on another exciting area " real time processing" which will gain momentum with applications moving to the cloud. There are broadly two reasons why this would happen:

1. Complex processes which needed large processing needs were relegated to batch mode of processing in the tradional infrastucture model will move to real time processing with lower to no constraints on processing power. Some powerful examples are "available to promise(ATP)" which will be available online/realtime if we kill multistage batch processing into single stage simultaneous processing of various areas like orders outstanding, inventory available, shop
Floor wip inventory etc.

2. Complex calculations and iterative methods needing large processing needs take multiple hours to complete the process. With the cloud, theoretically users should be charged the same for 100 machines running for 1 day or 1 machine running for 100 days and therefore you can process such needs with higher provisioning of capacity/processing needs for short intervals with no incremental costs.

Rapid elasticity and measured services (SLA based) can be attributed as two quintessential demand side requirements, amongst others, for cloud services to be successful. While, Cloud computing has miles to go & prove before enterprises completely adopt it, adoption is not very far away. It is a question of “When” and not “If”. Sample this:
• When a new application, Animito, was released by Facebook, there was an quantum jump in demand to bring the number of servers used on the Amazon Web Services from 50 to 3500 within three days ! This could not have been possible without relying on a cloud platform.
• The New York Times digitized and catalogued more than 100 years of archived articles for its Web site in a 24-hour period by using cloud offering
Enterprises also size their capacity for peak demand – An average corporate server runs at 20% capacity and 35% of network capacity is put to use. Adding the IT labour costs to perform routine activities these form the cost of doing business and do-not add to the organizational competitive advantage.
It would be interesting to discuss on how CIOs would look at these opportunities, what are the hard questions that they would need to ask ? What are the impediments to the adoption ?

One of the hard questions which would need a detailed answer would be "is this worth every penny of money spent"? How is this technology going to increase the inventory turns? How is this technology going to give me the competitive edge in terms of improvement in margin and reduction of lead time? And finally, one of the important questions: is this cloud going to be around long enough?
Security concerns, binding to legal regulations and questions about ownership of data still continue to be the nagging issues.

The awareness on cloud computing is self evolving and there will be a host of SIs to do the job. The real challenge would be in providing the consultancy on deciding for or against cloud computing. Organisations understand that cloud computing is more like taking the public transport system – Cost effective, fast, efficient and with a no. of alternatives and back-ups. To add to all that is the lower carbon footprint by using such public systems. There are the occasional hassles too like delays or inflexibility in routes and one takes that into consideration before one would hop onto a bus or a train. But still, a lot of people still prefer to travel by their own car.
I agree with Shruti and believe Abhishek provides some answers as well. It is famously said that 'if you cannot measure then you cannot manage' and the same holds true for cloud computing as well. Taking Shruti's point forward, I think, as a customer who has to decide on cloud computing, I would expect to be provided with an expert unbiased opinion on the parameters that will help me in taking the decision. Before cloud (or otherwise), we need to quantifiably answer why or why not cloud. As a SI we need to have a framework that will guide the customer in making the decision. There is a lot of peer pressure out there for organisations to move to the cloud and as is the wont of any new technology – there are successes and failures. The ones that are successful need to be studied and their context used to build the framework. Abhishek brings out an excellent start which can be further built upon by creating a multi-dimensional matrix of business and IT parameters that can go into building the framework. Based on customer's business needs, we could tag the infrastructure needs into a particular quadrant and then delve multi-dimensional for cloud specific parameters.

Check this interesting blog on cloud computing predictions for cios in 2011

http://blogs.forbes.com/ciocentral/2010/12/27/2011-cloud-computing-predictions-for-cios-and-business-technology-leaders/

Private vs Public Clouds: The benefits and leverage of the concept of cloud are different for public and private cloud models. All sources of variability are addressed by the public cloud while the private cloud only takes up a subset of all possible variability parameters. For example: 1. Cross Industry variability is not addressed by the Private cloud model.
2. Any kind of peaks/troughs on a particular day are not fully addressed by the private cloud model

While private clouds may not give the benefits of scale, they improve operational efficiency of managing data centers. Google, apparently,runs their data centers 50% more efficiently than most of the business enterprises.

To get the scale, enterprises may look at community clouds more actively. For e.g. Microsoft has set-up a gov cloud exclusively for governments

While Cloud offerings provide the opportunity to Consulting and IT organizations to expand their portfolio of services, the challenge remains the execution prowess. While the concept of cloud is still evolving, business will need guidance on what they would excatly need to do to leverage the potetantial of Cloud to stay ahead of competition. Here comes the role of the SIs to help them strategize their entire business around Cloud and then help them realize the benefit through deployment of a smooth cost effective execution model tailored to suit the business specific requirements.

Once the initial euphoria dies down, there will be a phase when organizations will evaluate the ROI on Cloud as well. Hence need of the hour for the SIs is to develop a Cloud Deployment methodology and framework, tool and differentiators, processes and templates to cash on this opportunity.