With 7 events all over the world in 2018 only (and counting), the second edition of the one-day-one-track conference taking place in Milan, 230 attendees (double than last year!) and 21 speakers from all over the world… we can say ServerlessDays Milan‘s got big numbers, and it even sold-out this year!
JeffConf turns into… ServerlessDays!
Formerly known as JeffConf (and if you’re wondering “Why Jeff?”, head to this article, which started the whole thing), ServerlessDays is becoming a bigger and more relevant serie of events time after time.
As media partner of the conference we went to Milan to meet this awesome community.
Fun facts & Welcome!
All participants are given an electronic name badge which truly represents the ultimate nerd gift (don’t believe it? Check it out!).
We also learn that every time a speaker runs out of time the horn of a Fiat-500-turned-into-live-tweet-screen will notify him/her that… time’s up!
The horn also serves as an energizer/”give a round of applause” for all speakers when they hit the stage! A nice and funny touch indeed.
Greeted by Alex, Luca and Simone, and an army of organisers and volunteers (thanks for all the hard work! You rock!), it’s time to start learning!
ServerlessDays Milan 2018: Here we go!
With a mixture of real-case studies, implementations and migrations to serverless, technical and less technical talks, and a couple of really inspiring “bigger picture” keynotes by Yan Cui (you can find the slides of his talk “You wouldn’t build your own toaster, would you?” here) and our trainer Gojko Adzic, there really was something for all tastes at this year’s conference.
On this blog post we’ll focus on the two keynotes.
But first of all, let’s not fall in the trap of objecting: “But there must be servers somewhere!” 🙂 Gojko Adzic summarises a right answer to this objection by stating: “Serverless is without servers in the same way Wi-Fi is wireless” in his blog post.
Would you build your own toaster?
Yan’s presentation kicks off by making a comparison between a company deciding to run serverless on its own infrastructures as opposed to “renting” the services of a vendor AND a person deciding to buy a toaster as opposed to… build its own!
The results? In the picture below!
Starting by this, there’s a few** key take aways on the world of serverless** (thanks Yan for summasiring them so neatly!) to be highlighted.
- Don’t pay for it if no-one uses it
- Don’t need to worry about scaling
- Don’t need to provision of manage servers.
These points imply a serie of changes in mind-set and open a number of possibilities that were simply impossible (or very costly!) before.
Deciding to host all of your activities on someone else’s servers (let’s call this “someone else” a vendor) opens up new possibilities. In fact, who decides to go serverless looks for:
- Cheaper costs
This, according to Yan, allows you to focus on creating value for the customer and therefore focus on what’s really important, rather than the pure technological aspects.
The shift needs to be from creating tech to creating value, something that serverless allows you to do.
Serverless’ recurring nightmare: The vendor lock-in
Another recurring topic of the day has been the so-called of “vendor lock-in“. Since what we would have normally hosted on servers is now hosted on, say, Amazon or Google’s infrastructure, we’re technically losing control over our data, website, etc. Isn’t this risky? What happens if Amazon or Google decide to shut down that very part of the infrastructure we’re being hosted in?
There were two main “schools of thought” here: Yan said losing control over the infrastructure is not necessarily a bad thing, considering the likes of Amazon and Goole have been investing into this increasingly (and decreasing costs for clients) in the past years.
Also, giving more control to someone else (provided this “someone else is a reliable company) means giving them more responsibility when it comes to maintenance, staff, offices hardware (someone must be working on it for sure, and those costs aren’t on you anymore).
On the other hand, Gojko, whose presentation “From universal software to superglue” close the event, has a more cautious approach, and believes “vendor lock-in” can be risky indeed.
Gojko’s shown data on the numbers of APIs that have been “killed” by the infrastructures’ providers over the past years (Google killed an astonishing 30% of them!).
Bottom line: be careful which provider you decide to use, even though, at the end of the day, your website will be most likely hosted on Amazon.
The Three Ages of Software Integration
As for Gojko’s presentation, he points out there have been three ages of Software Integration:
- Universal Software, when a software could you “anything you need as long as it’s configured”
- APIs, when everybody started publishing APIs, something which led to a lot of complexity
- Glue, where integration happens from the bottom up, and not from the top. This is the era we’re living in right now, and things are much more easy.
Referring back to the vendor lock-in issue, Mr Adzic shares with us the “Gojko’s Laws of Cloud Development“, stating you can only choose TWO among the below items from the list when it comes to Cloud Development:
You cannot pick all three of them. And this, he believes, it’s a good compromise anyway, since serverless allows you to be faster and do more.
To close with a quote from Simon Wardley, one of the leading thinkers when it comes to serverless:
“Serverless will fundamentally change how we build business around technology and how we code”.
Beer-Robots are taking us over!
And when the time comes to call it a day… who would have thought the ServerlessDays Milan organisers wanted to put such a nice touch to the end of this intense day of learning? They got a ROBOT to serve us beers!
So how was this second edition of ServerlessDays Milan? We steal the slide off Adam Matan’s presentation, who spoke in the afternoon, to say… ServerlessDays Milan SPACCA! 🙂 See you next time!