It was 1995 and in a garage in Seattle, three guys were about to launch their website and start taking orders for books. They ran the site on three Sun servers and they rigged up one of the servers to play a bell sound every time an order was placed. Within a few weeks the sound of the bell was constant and they turned it off. Amazon.com was on its way to success.
Five years later, writes SAP’s Fergus O’Reilly, at the beginning of the 2000s Amazon had close to 20 million customers and that garage with those three initial servers had grown to over 5 million square feet in fulfilment and datacentres worldwide with many thousands of servers.
Amazon could have just been happy to continue growing their e-commerce business at that point. But instead they saw the opportunity to invent an entirely new business, based on their existing infrastructure. Thus began Amazon Web Services (AWS).
Extending Amazon’s platform
AWS began by opening up the Amazon.com e-commerce platform with open programmatic access, open APIs. Third-party developers could now build software to tap into the Amazon product catalogue, check availability, access customer-created content like ratings and reviews, use the search engine, the shopping cart and the famous Amazon 1-Click purchasing system.
This allowed other companies to build their own extensions to the Amazon platform, building their own selling sites powered by AWS under the hood. As Amazon said, “they surprise us by building features we never thought of.”
Within a few years, over 100,000 developers had signed up with Amazon to use the AWS platform in their own e-commerce applications.
is Chief Solution Expert for SAP Consume to Cash
Amazon then saw another opportunity. By this time they had 50 million customer accounts and deep expertise in managing massively scalable, servicesoriented IT infrastructure in their data and fulfilment centres. This core competence was focused on serving their internal needs. And there was a relentless focus on datacentre efficiency so that operational savings could be passed on to customers in the form of lower prices to gain competitive advantage.
… And into the cloud
Amazon then realised that other companies could also benefit from what they had developed. They launched compute and storage capacity in the cloud offerings as part of AWS.
This allowed developers to run their applications and store their data in the Amazon cloud. Developers could just concentrate on building applications without having to do the grunt work of building and managing scalable datacentres. Amazon monetised this with a pay-as-you-go, usage-based revenue model.
In doing this Amazon had essentially turned their systems inside out. Rather than just using their world class infrastructure for their own needs, they found a way to open it up so that others could consume and use it also.
They rapidly grew this part of their business, and over half a million developers now use AWS services. The compute capacity available in the Amazon cloud is now estimated to run on well over 40,000 servers, probably an order of magnitude larger than that which Amazon uses just for running its own commerce business.
Amazon does not report annual revenues from their cloud business but they are estimated by analysts to be over US$500 million today, and growing quickly to reach $1 billion within the next year or so.
A roadmap for M2M
This Amazon cloud success story should serve as an inspiration and even a roadmap to success for many companies entering the machine to machine space. In particular, telecom operators are waking up to the fact that they too need to turn their businesses inside out in order to address the opportunity in M2M.
Operators, like Amazon, know how to runmassively scalable datacentres: large telecoms companies run some of the biggest datacentres in the world. Operators are also experts in providing network connectivity to tens or hundreds of millions of disparate end-point devices: the phones, tablets, TVs and consumer electronics that we all have in our homes, workplaces and pockets.
Operators know how to mass provision services to those devices. They also know how to diagnose and fix faults across their services and networks, and how to remotely manage and update the endpoint devices connected to their network.
Finally, operators know how to track the location, status and usage patterns on those devices, reliably bringing that information all together in their billing systems to aggregate billions of micro-transactions to generate revenue.
Today these core competencies are all turned inward: telcos think about how they can efficiently run their datacentre, their network, and their services to provision, track and manage devices that they sell, to collect revenues that their customers pay them directly.
Can telcos turn inside out?
But M2M businesses also need to deploy and manage large numbers of smart devices across a wide geographic area, connect those devices back to applications in a datacentre, run services distributed across the network, track usage patterns, bill customers based on device and network utilisation, share revenue with partners and constantly analyse customer needs to keep ahead of the market.
It is evident that much of what any M2M business needs, telcos already do.
The inside out play for telcos means that they need to transform just like Amazon. Figure out how to take their internally-focused capabilities, and turn them into open platform services that developers at other companies can easily use to build their own M2M businesses.
This is not an easy transition. And running the combined businesses of many M2M customers on a single, shared platform means scaling that platform by one or two orders of magnitude from where it is today
Succeeding in this business requires rethinking many of the core processes at the heart of a telco. Business processes must be designed to operate with minimal human touch points, since that is the only way to scale. M2M companies need to be able to come to a telco and start leveraging the telco platform infrastructure on demand, without having to go through long, tortuous sign-up processes or wait months for service to be turned on. Platform capabilities need to be easy to access, operate reliably and come with clear service-level agreements and transparent pricing schemes.
In many ways this inside out platform play should be a natural ‘homerun’ for telco operators. It can be a massive driver of revenue growth over the medium to long term. The business is theirs to lose. But if they do not move rapidly to capitalise on the potential then they will be bypassed.
Other players such as cloud computing providers like Amazon, Google and the major IT and SI companies could move rapidly into providing such turnkey M2M platforms. These players do not today have the same capabilities as telcos when it comes to such things as communications networks, remote management of smart devices and operation of efficient micro-transaction billing systems and high-volume analytics. But they could all take steps to acquire these for themselves.
The clock’s running
If no one moves fast enough to become a complete and global M2M service provider then those looking to build new M2M businesses like companies in the manufacturing, utilities and hightech industries will develop everything themselves: building their own datacentres, buying network capacity, figuring out how to mass provision and manage remote smart devices and how to analyse and monetise transactions across those hyperconnected devices to generate revenue and share it with partners.
Smart companies who build this will then inevitably realise that they can become M2M service providers themselves, using the platform they built not just to run their own business but selling its capabilities to others with similar needs.
Amazon showed the world how to successfully turn your core IT systems inside out, it is an approach just waiting to be emulated.