Career 5: My very first patent - a memorable fond recall (v1.0)

 I worked in the HA (High Availability) labs of a company. Initially I worked on failover scripts of single points of failures in some of the mid-tier middleware products. Failover is used for any single point of failure and most commonly applied to databases. I was the only person in the lab with middleware and mid-tier expertise.

The lab manager of the lab was concerned by a problem. When a database fails over, unfortunately clients will see a disruption. We also do not know the outcome of work in progress - did they complete or not? It is very difficult to determine the outcome of these after failover, sometimes even manually. Determining the outcome is a key step before handling the recovery. This poses a big and costly problem for businesses. Say you deposited funds using an ATM. After a failover while processing it, there is no way to know if the deposit went through. Even with manually checking, you need to know the previous account balance before you know if the deposit took effect. Even then there is no telling if somebody else slipped in and made a change!! He threw out a challenge to the whole lab to find a way to solve this problem. This is a hard problem. 

The only one who came up with a potential solution was me. I leveraged a three-tier architecture for the solution (almost all modern systems are 3 or 4 tiers consisting of at-least a client, an application server and a database). The idea was to automate the retry triggering and outcome determination in middleware from these so no manual actions are required. The actual client does not even see a disruption during the failover. I partnered with a team at the company's prestigious well-known corporate research labs, to prototype my solution idea using a CORBA mid-tier. We fleshed out more of the details and wrote up and applied and got a patent on this with the US patent office. I don't know of any other patent in this area. The corporate labs continued on to break further ground in this area by applying and getting two more patents building on this foundation. 

This was my very first patent!! Your first patent sticks in your mind for your lifetime. Your first patent has a special place in your heart. 

I moved on in the HA lab to a project to achieve extremely high availability (the so called 5 nines project). My focus was low latency disaster failover.  I however was frustrated because as a database and middleware guy, in my view, this was not a realistic goal when the company did not control the database and mid tiers. There were also some conflicts within the team. I also concluded that my skill set was increasingly mismatched with the company's direction. I left the company and joined a high-flying internet company as the platform/infrastructure architect.