The Archaeologist Who Went Serverless (and why you should too…)

There’s a lot of hype now about ‘serverless’. Almost every day we read about new serverless application frameworks, serverless cloud providers and, of-course, serverless t-shirts.

First, a little background, in case you’ve been living in a log cabin in the woods and are unfamiliar with the world of serverless. (And actually, we’d be ok with that living arrangement, as long as the cabin had wifi, a beer-stocked fridge, and a fireplace. But we digress)  The term ‘serverless’ is a bit of a misnomer, like white ‘chocolate’ or ‘reality’ television, – it doesn’t mean ‘no’ server, it just means ‘someone else’s server.’  A lot of things are called ‘serverless’, but all of them include a server that is operated by someone else. So there is still a server, but you don’t have to waste your time, effort, and money dealing with it. And for a coding team under pressure, this can save a lot of time.

Here at The Research Software Company, we were approached by an archaeologist from a major university’s research lab, and asked for a web application that performs queries on a data set. She has very carefully built a unique data set over the course of a decade, and she wants to give other researchers access to it.

The data set is large archaeologically (spanning almost all the available data of its kind), but it’s also relatively small viewed through the eyes of us data nerds – only a few megabytes in total. We often find this to be true in Digital Humanities research that is very specific.

All web applications that let you query a database look pretty much the same – there’s a frontend in some Javascript dialect and some framework, and a backend with a database behind it (We use Python and Postgres if we start from scratch).

In this case, since the database is so small, we decided to do something else – we will not implement any backend. We’ll implement all the queries in the frontend. The frontend will just need access to the data set, which can be quite neatly stored in a single CSV file – so there’s no backend!

We explained to the archaeologist that she doesn’t need to live in the past and rent a server and a database. She will only need static storage for the frontend files and the CSV file and that’s it. For all practical purposes, she really is serverless. Yeah, we know  technically there will be some web server delivering all those static files. But our friend the archaeologist and her staff won’t have to lose sleep worrying about it. So in this case, they are using the technology of the future to give them more time to spend studying the past.


Leave a Comment