Become a Google Cloud Platform Expert
By Raymond Blum, Author (GCP in action)
There are a number of personal missions that I set out to pay service to on every project and every team that I work with. One of these is to remove the hubris of software development as an industry, in the belief that we have invented many of our practices and patterns. This belief is often a myth when in fact we have adapted solutions that were tried and true in business long before the existence of digital computing technology.
A key strategy employed in my recently released book "GCP in Action" is to bring such "prior art" in computing and non-computer business practices to light for Software Engineers and System Architects.
This is reflected in a discussion of a specific technology feature that is utilized heavily in my book. The concept of "Eventual Consistency" was popularized in the Internet Domain Name System and then in mobile computing and NoSQL databases, including most famously in Bigtable.
A system that guarantees Eventual Consistency allows for temporary data discrepancies between replicas, guaranteeing that all copies will eventually update to the same value. It prioritizes high availability and lower latency.
While this is seen within computing circles as originating in the above implementations, it is actually the recognition of the project management principle that recognizes Urgency and Importance as separate aspects of a milestone or achievement.
For an achievement to be "Important" means that there is a significant negative effect on the value of the final product if the achievement is not made. For example if our accounting system has a feature named "generate tax return feeds," it is important that this feature be working at the end of a tax reporting period such as April 15th in the US. If our customers were not able to rely on this aid in filing their returns, they'd definitely find our system to be less valuable.
On the other hand, urgency is recognition of the loss of value if an achievement is not made by some arbitrary and close point in time. Staying with the tax filing example, our system is not significantly less valuable if the "generate tax return feeds" feature is complete and working on April 8th or February 2nd as opposed to January 20th. On January 20th the feature is not urgent, nor is it on February 1st. When April 7th comes around, the feature is more urgent and on April 14th, the urgency is critical. The feature was always important but urgency was not constant over time.
The example application in "GCP in Action" considers the tallying of all online votes to be important but not urgent until the results of the vote are needed. What is urgent is confirmation of recording each vote when a voter casts it.
The most famous illustration of this separation of urgency and importance is known as "The Eisenhower Principle." Dwight D Eisenhower, allied commander in World War II, is quoted as saying "I have two kinds of problems, the urgent and the important. The urgent are not important, and the important are never urgent." Eisenhower did not claim this insight for his own, but attributed it to an unnamed "former college president."
This time management technique of project management guaranteed that important things are not forgotten and that non-urgent tasks did not starve those which are truly urgent from getting attention. This quote and the manual method used to implement it, known as an Eisenhower Matrix, precedes the use of computers in business by decades.
As further examples of computing protocols taken from the analog world, consider the wide use of consensus algorithms such as a Paxos algorithm.
To see the analog origin of this, go to your favorite streaming provider and watch the 1995 movie "Apollo 13." In the movie there is a dramatic scene in which banks of NASA technicians are seen using slide rules to perform manual calculations, recording the results on paper. The results are then passed up a hierarchy and confirmed with redundant counterparts and only if no mismatch is encountered, is the result accepted.
While the specific application of slide rules in the film is not technically accurate, the process of requiring a quorum of consistent results from redundant independent workers was. This procedure was used in applications where the accuracy of manual calculations was critical. The digital implementations of this were not born in a vacuum, but rather adapted from established practices.
The strategy of researching, recognizing and adapting existing solutions is key to the most effective use of a rich infrastructure platform such as GCP.
