Scalability in Cloud-based Solutions: Why is it needed?

By Subhayu Roy, Co-Founder, Notebook

It was April of 1996. Yahoo got listed on Nasdaq, and more than doubled in value the very first day of trading. It was a significant time as it signalled the arrival of the dotcom era. Faith in online businesses has shaped business and human history for the decades since.

Imagine India in 1996. Think desktops powered by Intel Pentium I (or Pentium MMX if you got lucky). The first mobile phone call is less than a year old. Relatives coming from Dubai are bringing you calculator watches. Now imagine yourself to be a tech entrepreneur in India in this era. How many users would you plan to serve? According to VSNL, who have launched the GIAS (9.6 kbps at Rs. 25,000 for 250 hours) in India on August 15, 1995, India has a little over 10,000 internet users in 1996. How would you store the data to serve these 10,000 users? The term ‘Client-Server’ has already been bandied about since 1991, so you need to set up a ‘server’ and for this, you need a ‘mainframe’. Moreover, your accountant would advise you to factor in air-conditioning, electricity and additional manpower costs for the upkeep of this wondrous system.

Scalability in Cloud-based Solutions

Fast forward to 2019. Following some tumultuous times, Yahoo has been acquired by Verizon 3 years ago. Your mobile phone now boasts of a memory larger than the 1996 server you had installed. Imagine you are still a tech entrepreneur. The 566 million internet users in India now form your addressable audience. Internet is now available at 4G speeds on your handheld device at a rate lower than anywhere else in the world. More than 100,000 village panchayats have been connected with optic fibre for high speed internet access. Around the globe, similar developments have taken place. The next Billion users are waiting for you to serve them. How big of a ‘mainframe’ are you thinking?

Why did we need to build for scale?

Subhayu Roy, Co-Founder, Notebook

As co-founder of an edtech startup in India, it was clear right from inception that the only way to meet our objective was to achieve scale. India has more than 250 million school students across 1.4 million schools. Current edtech products either don’t reach a vast majority of them — being offered in physical forms like preloaded tablets or flash drives — or fall outside their economic reach. We knew we had to offer online streaming as a delivery mechanism, and develop an innovative commercial model to drive adoption. Once the student gets access and tries the product, we trust our content to do the rest. The economics of this approach only works out when there are enough students to offset the content creation costs.

But every business wants to scale. Where is the problem?

The issue with scaling in India poses a special problem. The economic diversity translates to a situation, where the moment business scales beyond the top urban-elite customers and heads towards the next Billion, the ability to pay decreases considerably.

Imagine a Supermarket with one till. As business grows, it puts in more tills, because every incremental growth in number of buyers means an equivalent growth in revenue and profits. So, costs grow proportionate to revenue. However, if every successive buyer buys a little lesser that the previous buyer, the ROI on every successive till decreases. Hence, scale affects the business adversely. The supermarket either has to absorb the capital expenditure on additional tills or stop buying tills and allow the customer to wait in long queues.

In our case, serving students in the farthest reaches of the country with the same quality and experience as those in metro cities was a non-negotiable business objective. Adding the infrastructure in the anticipation of these numbers, though, was well outside our reach. Scalability in Cloud-based solutions like storage and delivery, to us, provided the ideal solution. It allowed us to manage our expenditure on IT infrastructure in line with our business growth instead of laddering up in steep increments.

What are the Cloud Computing Advantage?

Technology is ubiquitously interwoven into our lives. With the ever-expanding user base, the sky (or should we say the infinite realms of human imagination and capability) is the limit for Cloud Computing.

We are witnessing a massive growth in the adoption and evolution of cloud computing. Cloud providers like Microsoft, AWS, Google and others have spearheaded a paradigm shift in the way organisations envisioned their IT infrastructure, and have themselves emerged as major PaaS (Platform as a Service) businesses.

With the advent of Virtualization, as a logical step, PaaS has become the next game changer. The Cloud Service Providers biggest value-add is to offer an easy-to-manage, easy-to-scale service or infrastructure layer which can dynamically scale to match requirements. This enables startups and “value-builders” to focus on the value proposition they want to take to the market, and core technology of the solution they have envisioned. The DevOps challenge is now standardized and small companies can own large servers at times of need and pay only for what they use.

Gone are the days when the lure of a data centre or expensive servers meant an obnoxious CapEx investment – which would be an entry barrier. As times have changed, Cloud has brought expensive infrastructure and scaling capabilities to a PAYG (Pay As You Go) model. The future of programming, as envisioned by the Cloud Service providers, is “Serverless programming.” Serverless allows you to deploy your code in an infrastructure and not be bothered. It’s ready and built to scale and supports the load you throw at it.

Managed Databases, CDN Services with global POP, Queue Management Services, Mail Servers built to Scale, Load Balancers, Layer 7, AI and ML tools, Speech to text and vice versa, Bot Management, AI and ML based Platforms, Media Services and more are available on the cloud.

 Why do business need Cloud Computing?

Uncoupling business aggression from capacity building – Any new business strives to achieve the maximum impact within the shortest time. This, in today’s world, means sizeable advertising spends, PR exercises, on-ground activations, trial offers along with freemium offers, schemes and discounts, to drive adoption. For every customer acquired, if the business also has to keep a watch on capacity building (which can only happen in large increments), it creates a tightrope walk for the business and puts massive amount of strain on its resources. Secure in the knowledge that capacity will match the demand, thanks to a Cloud-based model, businesses can now pull out all stops in planning their growth.

Freeing up resources for developing core business – Back in the days of mainframe, every business will have to set aside a monthly manpower and overheads budget for the upkeep and maintenance of the infrastructure. Accounting for outages, delays and other unforeseen eventualities, every business would end up setting aside a large chunk of its spends that could have been utilised on resources critical to the core business value proposition. Licenses would also take up some resources. Texas-based solutions company, BMC, conducted a Mainframe Survey in 2015 which found that 48% of the companies surveyed spent between 25% and 50% of their mainframe software budget on monthly license charges (MLC) and another 22% said that they spent more than half of their mainframe software budget on MLC costs. When these costs are scaled in accordance with business growth and customer acquisition, it enables the business to employ better resources to deliver a better product, service or solution.

Add on services for seamless delivery – Competition in the Cloud space has driven every provider to develop a comprehensive layer of services on top of their storage facility. While designing our architecture, initially we had looked at how we might have to coordinate with multiple providers for different pieces of the solution to deliver a seamless video streaming experience. We have been pleasantly surprised to find services to help us store, transcode, protect and deliver the videos in the most efficient manner. You also end up saving a considerable amount of development time since these services come pre-integrated with the storage layer.

Ease of managing distributed teams – This was perhaps the earliest sales pitch for the Cloud: the “anywhere, anytime” promise. For us, having data on the cloud has enabled multiple experts from various geographies, across multiple time zones, to operate as one cohesive unit to deliver a great product.

Coming back to our example of the Supermarket with one till, Cloud is like a service that offers a bunch of people with POS machines. These people come in during rush hours, and ease the billing rush. Given that a few such agencies are trying to win the Supermarket’s business, they also provide people to bag the goods, carts to move the goods, and sometimes even drop your purchases home!

A few days ago, we celebrated 30 years of the World Wide Web, first proposed by Tim Berners Lee on 11^th March, 1989. The internet was designed as a global, larger version of the US Military’s ARPANet developed in the 60s. The primary consideration behind this was Disaster Recovery Management. Incidentally though, at its core, Cloud Computing is the same distributed computing philosophy, now applied to most online businesses.

At Notebook, getting on the Cloud is us keeping our promise to offer the same seamless product to every student, irrespective of the city, town or village where the student is studying from. It gives us the ability and hence, the confidence, to aspire to touch the lives of the next Billion internet users, to educate a whole generation of leaders.