It is hard to believe now, but when I joined Confirmit in 2001, the terms “Software as a Service” and “On-Demand” didn’t really exist. They are such a core part of who we are and what we do now, that it barely seems possible that Confirmit fell into the world of “Application Service Providers” or ASP. It also underlines exactly how far ahead of our time we were in terms of our technology and business model. We were offering multi-tenant SaaS long before the term was even coined.
No one else was doing anything like that. But then again, no one else was creating software like we were either.
Although being such trailblazers sounds rather glamorous and exciting, the reality was that we had a lot of very manual checks to do to ensure we kept everything running smoothly. As an example, during the time I started working for Confirmit, one of the more important operational routines was to check that the ASP server was running when you got in to work in the office every morning. Now, not only do we have constant visibility into every aspect of the SaaS environment at all times, high-availability architecture, and systems that can to some degree automatically correct issues as they arise, but it is also under the vigilant eyes of our partners at Rackspace.
In 2001, the “ASP server” was actually a standard desktop computer that was also being used by one of our developers to make changes to production code. From tiny acorns…
Obviously, as customers began to see the benefits of our approach, it became clear very quickly that we had to move to a much more comprehensive and scalable hosting model. At that point, “comprehensive” meant a four-server system hosted with our IT partner in Oslo. This meant we could start modularizing the hosting environment to provide much better performance, but we still did not support any load balancing or redundancy features. Both of which were still a way into the future for the industry, not just Confirmit! All customer data was backed up, though, with the company’s receptionist dropping off tape backups to an offsite location.
Things moved fast, though, and already by October 2001, we finally got Confirmit running on proper infrastructure. I was instrumental in the data migration - flying to New York with all the SaaS production data on a NAS in my backpack, and delivering it in person from a yellow cab to the AT&T datacenter in downtown Manhattan… It may sound crazy now, but back then a full 250GB of compressed data would have taken longer to transfer over the Internet to the new servers than the flight across the Atlantic took me. At the time, the biggest considered risks were the recently introduced restrictions on airline traffic, and the possibility that DHS might mistake the NAS for something more sinister, but fortunately, we did not run into any issues. AT&T techs connected the NAS to a server in the data center, and our people in Oslo, who had been on stand-by since my arrival at Newark airport, started on restoring the data while I got ready for the return flight to Oslo.
We have come a long way since then…
Digging through some old documentation folders, I was able to find this diagram showing the site layout as of July 2001:
Compared with today's diagrams for just one of our sites, it is plain to see there has been more than a few changes.
Now, with two consecutive years of 100% uptime in 2014 and 2015, and last year’s launch of the environment in Australia, there is definitely a lot more to keep track of. Our management of SaaS and operations in general has matured in a big way in 15 years. I cannot even imagine what it will look like 15 years from now, but I am excited to find out!