Creating a Startup on EC2 – Really a Good Idea?


(Foto: Leonard John Matthews)
(Foto: Leonard John Matthews)

Amazon EC2 – A New Platform

Amazon is one of the pioneers of providing infrastructure as a service (Iaas). Customers can instantiate virtual machines via the Web with complete control over all computing resources and pay only for the resources they consume: computer power, storage and network  bandwidth usage. Such an environment allows to scale up and down computing environments in minutes thus enabling new and potentially more efficient ways of running datacenters and IT services. Amazon created a whole new platform (comparable with Microsofts Windows Operating system) that generates a virtually illimited  number of opportunies for startups to provide new services, tools, and applications to build on top or to complete Amazon’s offerings.

This is the optimistic version. But let’s dig deeper – is it really a good idea for startups to build upon the EC2 platform? What are possible strategies to be successful? What are the risks to fail?

Platform Software

Becoming the owner of a popular platform, i.e. a foundation product that enables other software companies to build own offers upon, is the ultimate dream of every software company and its investors. Platform owners define the rules of the game and profit directly and indirectly from any economical activity on their platform since every new product based on the platform increases its attractiveness. IBM’s Mainfraim systems, DEC’s Minicomputer Operating Systems, Oracle Databases, Intel’s 386 Family, Microsoft’s Window OS, Java – all those are examples for successful  platforms that became de-facto standards. The Internet itself is such a huge platform that it has become the foundation of many other platforms like EC2, Google’s App Engine,, but also Facebook or Twitter.

The downside of the ambition being a platform provider is the extremely tough competition. There’s typically only space for one or two platforms in a segment that defines the standard. The way to get there is challenging, expensive and extremely risky. Only a few can win.

Succeeding as Niche Provider

It is far easier and more predictable for small companies to position themselves as provider of complementary products for a given platform (like building Management Applications for Oracle Databases, Debugging Tools for Windows, Java Optimization Software, etc). Occupying and defending a niche has shown to be a successful strategy for startups many times. Some of those startups even expanded from their niche into horizontal markets and then became themself platform providers (e.g. Microsoft, Apple, Google). Others developed outstanding and unique skills that attracted the interest of the platform provider that eventually bought the startup and integrated their technology in their own plaform offererings. Being bought by Google is probably the most popular exit strategy of web-startups today. Google bought more than 50 companies between 2001 and 2008, Microsoft even more.

Is there any evidence that this strategy works also for EC2? Amazon acquired very few companies, almost all in the retailing market, one UI-specialist (Shelfari) and one stake in an alternative payment company ( . There is no company in the list that offers technology to improve, extend, or complete the EC2 platform. Does Amazon at least leave enough space in its eco-system for niche and complementary providers?

Looking at the history of EC2, it is quite impressing how the platform evolved within short time intervals. EC2 started as beta-version in 8/2006 without any graphical interface. Access to the platform was only possible via an API. There was no possibility to assign persistent storage to running instances, neither static IP addresses. In 12/2007, they presented DevPay (only for the US), a payment system that allows providers of virtual appliances (AMIs in Amazon’s terms) to assign fees to their VAs that Amazon charges from the users and credits to the provider. In 3/2008, they came out with features for high availability like static IP addresses and availability zones. In 8/2008, they provided the possibility to attach persistent storage to instances. In 10/2008, EC2 lost its beta-label and went into production mode. Two months later, EC2 become available in Europe.  Beginning 2009, Amazon announced a management console.  Shortly later, they started providing images based on Windows. In 3/2009, it became possible to reserve instances, i.e. keep fully configured images in hibernate mode without paying for them until they are used. And some weeks ago, they added monitoring, autoscaling, and loadbalancing to the platform’s features.

Given this impressive list of innovations, it is clear that Amazon puts a lot of pressure on startups to remain competitive against the platform provider. This brings us back to our initial question: is selling products or services on for the EC2-platform really a viable business model?

Strategies to Survive

We believe that this is the case. We even believe that the situation is not substantially different from any software provider which builds products for Windows. But any startup should be extremely careful concerning their decisions on the product/service scope and feature set. Here are a couple of strategies that may work:

  • Startups could provide solutions for a technological niche that is not of common interest of the vast majority of EC2 users (e.g. network simulation)
  • Startups could concentrate on specific usecases for a subset of potential EC2 users and provide simple solutions to streamline those usecases (e.g. deployment of web-services in Ruby on Rails)
  • Startups could focus on a vertical industry segment (e.g. provide solutions for lawyers)
  • Startups could diversify their offering by supporting different infrastructure providers (like GoGrid or FlexiScale)
  • Startups could try to be faster than EC2, sell solutions for existing gaps in the Amazon offerings, and move on when EC2 fills the gap itself (RightScale seems to follow such a strategy)

Whatever strategy a startup chooses, an ermerging growth market as cloud-computing and infrastructure services should offer enough possibilities for determined founders to succeed and create a thriving business.

2 thoughts on “Creating a Startup on EC2 – Really a Good Idea?

  1. Pingback: Twitted by nicholasdr

  2. Pingback: How will Enterprises manage their Infrastructure in the Cloud? « Elastic Security

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s