How a Camera Company Started The Serverless Revolution (Then Killed It)

We are in the midst of the serverless revolution. Everything around us is going serverless. We have serverless functions, serverless databases, serverless ETL software, and I could go on and on about all the things that are serverless or perhaps going to be serverless sometime in the future. But, things could have been a lot different. The serverless landscape that we see today could have been much different if a particular camera company had not underestimated the power of utility computing.

Fotango 1999

It all started back in 1999 when a UK based software company emerged from the depths of the internet. The company went by the name Fotango.

Fotango was a photo application that had a rich set of features. Its business niche was at the intersection of photo processing and online photo sharing and storage. Fotango visitors would register with the platform and login to request a prepaid postage envelope. They would put raw photo strips into the envelope and mail them to Fotango for processing. Fotango would create physical prints from those photo strips and send those prints back to the user while storing digital copies on the platform. The user could view those digital copies in their Fotango account and even share them online.

In the year 2000, the dot-com bubble prompted large companies to expand to the digital space. Cannon was shopping around for a company to acquire to help its expansion. When they learned about Fotango and Fotango's market position and trajectory, they decided to buy them. By 2002 Fotango became Canons child company.

Zimki 2005–2006

In the year 2005, Fotango's research and development department launched an experimental platform for application developers called Zimki. It was a first of its kind end-to-end javascript solution that enabled programmers to write backend and frontend of their application in a single programming language, javascript. The platform conveniently abstracted away infrastructure complexities so programmers did not have to allocate efforts towards that stuff nor did they have to focus on building backend API's nor provisioning, configuring, and maintaining servers. They could direct their attention towards the development of their or their client's product. Several elements of this concept overlap with the serverless concept we have today. The common thread is the application infrastructure has been offloaded to the vendor, so the application programmer does not have to bear the responsibility of scaling it.

Zimki's backend abstraction became an attractive feature to their target audience and made it the go-to solution to get applications off the ground without incurring substantial development costs. The platform provided so much abstraction that programmers could launch an application in just a few hours instead of days.
The platform also became the first of its kind to implement elastic pay per usage payment model. Zimki customers would only pay for the resources their application would consume. There were three pricing meters:

  • Bandwidth
  • Storage
  • Operations in the virtual machine

If their application did not consume any resources then they would not have to pay for anything. This was the first “pay per compute” pricing model that hit the technology scene.
Basically, the more of the resources you consume, the more you pay — and of course, vice versa. — James Duncan

Zimki 2007

In the year 2007, Zimki started to have good traction. They were one of the largest hosting elastic IAAS (infrastructure as a service) platforms for developers. Everything seemed to be going well for Zimki. They had a lot of money, Fotango backed them, and pretty much the future looked bright. Unfortunately, things were going to change.

The Zimki team started to work on strengthening their authority in the market by tackling their prospects biggest fear and complaint, vendor lock-in. To fix their vendor lock-in issue, they made a strategic decision to open-source the Zimki compute engine, and they would position themselves as a centralized hub for all Zimki services. By open sourcing their engine they would stimulate competition between hosting providers. Hosting providers would offer their platform as a service which would give Zimki developers the freedom to move from one hosting provider to another, thus eliminating vendor lock-in issue. Zimki programmers would also have the option to host their version of Zimki on their servers if they wanted to.

The system was going to be GPLv3'd precisely because of the “SaaS” loophole. The intention was to allow multiple providers to make operational improvements to the code base for reasons of service competition whilst ensuring portability through a separately established assurance authority with a trademarked compliance stamp. As part of this, an exchange was to be established. — Simon Wardley

End of Zimki

Fotango started pouring marketing dollars into that Zimki’s open source initiatives; branding the platform as the utility platform of the future. The idea that the platform was going to be open sourced helped grow Zimkis user base. They went as far as becoming platinum sponsors alongside with Google, Microsoft, and Intel for the OSCON 2007 conference to announce they are going open source. OSCON 2007 was of one of the biggest technology conferences in the world at that time.

Unfortunately, the open source announcement did not go through well. Canon learned about Zimkis initiatives and demanded the compute engine not be open sourced. They protested Zimki because it was not aligned with their core business values and ceased all funding for Zimkis parent company Fotango. This protest happened right before the big OSCON conference. James Duncan (CO - CEO/COO) proceeded to make the announcement anyway because that was the main objective of the conference and it was the leading motive for why users signed up with Zimki in the first place. Simon Wardley, the CEO/COO of Fotango resigned right after James Duncan made the announcement. Zimki slowly ceased all operations and had shut down at the end of the year.

If Zimki had been still operational today, the platform could have been one of the most adopted, serverless platforms out there. The platform would have an influence on the evolving serverless landscape. Canon could have been a major competitor in the cloud industry, competing against Amazon Webs Services, Microsoft, and even Google Cloud.