Cloud Next 2018— a whirlwind week at Google’s annual cloud conference

Somewhere downtown in San Francisco

The last time I attended Google’s annual cloud conference was in 2014. Back then it was actually called something completely different, ran for just one day, and had just 250 (invite only) attendees. Ah, the good ‘auld days of yonder years!

Anyway, this year the it spanned a whopping 3 days (excluding other periphery events like the Partner and Community Summits), and the estimated number of attendees was 20,000. Yeah, I think it’s safe to say that this Google Cloud thing is kind of a big deal now!

In this article, I’d like to share my highlights and key takeaways of Cloud Next this year. I’m not going to deep-dive into topics because I’m far too jet lagged and lazy to do that. Instead, you’ll have to wait for my brain to start working again so I can write some decent tech articles.

Note: you can find recordings of all the keynotes and breakout sessions here.


Community Influencer Summit

Melody Meckfessel, VP Engineering on GCP

This year I was invited to a community influencer summit on the day before the actual Cloud Next conference started. In fact, Google were incredibly generous and covered all my travel expenses for the entire week through this invite. So, thank you for that guys!

This was one of my favourite days of the week. It was an intimate gathering of Google’s key people who drive their community program, and people who they recognise as the leaders in the developer communities. It felt more like a user group meetup, rather than any kind of summit. I loved that about it.

We were treated to some great talks, in particular one from Google’s David Rensin (who quite literally wrote the book on SRE by the way), and when he opened his talk with this, I just knew it was going to be good:

“I’m an engineer. I don’t do slides. Just ask me questions and we’ll take it from there.” — David Rensin, Google

I also got to catch up with some of my fellow GDEs that I hadn’t seen in years. This was fantastic. In fact, I got more value from chatting with some of them than I did from the actual breakout sessions during the conference itself.

Servian’s strong presence

Andy representing!

Servian (the company I work for) had a big presence at this year’s event. In total, 7 of us travelled to the conference. Servian are an official Google partner and doing some amazing things on GCP — in fact, we’re the only one in Australia to be given the Partner Specialisation for Data & Analytics award.

Apart from spending some quality time with my Servian teammates — many of whom I’d never actually met in person before — one of the key moments of Servian's involvement during the week was Andrew Pym speaking on stage during the Partner Summit. He did a great job representing Servian on the big stage!

Day 2 keynote — BigQuery’s time to shine

You don’t even need to drop out of SQL anymore to build models — boom!

In total there were 105 announcements from Cloud Next ’18. It would be folly of me to try and list them all here or to talk about them in depth. Anyway, the clever marketing folks over at Google have already compiled a complete list of them for your perusal here.

Now, as many of you already know, I‘m a bit of a BigQuery fanboy. In my opinion it’s the best tool on the stack, and the gateway drug for many to GCP. On day 2 during the keynote, we were treated to a torrent of new BigQuery features .

The top 3 for me were:

1. BigQuery ML (beta) — a new capability that allows data analysts and data scientists to easily build machine learning models directly from BigQuery with simple SQL commands, making machine learning more accessible to all.

2. BigQuery Clustering (beta) — creates clustered tables in BigQuery as an added layer of data optimisation to accelerate query performance.

3. BigQuery GIS (public alpha) — new functions and data types in BigQuery that follow the SQL/MM Spatial standard. Handy for PostGIS users and anyone already doing geospatial analysis in SQL.

Kelsey Hightower’s dope keynote!

“SQL is dope!”

It’s not every day that you get to see Kelsey Hightower live in action. So, when the keynote for day 3 rolled around with his name on the agenda, I made sure to get there early to secure a good seat! #fanboy #stalker

This man is seriously talented. In front of thousands of people, I watched him ad-lib his entire keynote, copy and paste from Stack Overflow, and code a bottom up solution in 15 mins on GCP! It was definitely the most useful part of the week for me. He also quipped some absolute nuggets of gold during his session:

“SQL is dope!” — Kelsey Hightower, Google
“Remember, the price of admission to Kubernetes is a container image” — Kelsey Hightower, Google
“Good programmers copy; GREAT programmers paste” — Kelsey Hightower, Google

Update: There’s now a YouTube video of just Kelsey’s demo here.

GKE on-prem

Google’s new ‘Cloud Services Platform’

The K8s steam locomotive just keeps on rolling folks! There were lots of announcements around GKE/K8s, but none more interesting than the one about GKE will soon be available to be run on-prem. In essence, it’s managed Kubernetes on-prem. I think the official docs do a good job of describing it, and it’s far better than any drivel I could write:

“With GKE On-Prem, you get the Google Kubernetes Engine (GKE) experience directly in your data center. A quick and simple install and upgrade experience that’s validated and tested by Google. GKE On-Prem also registers your cluster with Google Cloud Console in order to have a single-pane-of-glass view for managing all your clusters. The abstraction layers built into Google Kubernetes Engine and GKE On-Prem allow you to make your applications “cloud-ready” and move them to the cloud at your own pace, all while developing operational skills that are portable.”

It’s currently early access only, so you’ll need to sign up to get playing with it. This move underscores Google’s commitment to having hybrid/multi-cloud configurations, which will also keep the big enterprises very happy. Good move Google.

Serverless containers

This is a game changer!

Wow. Just wow!

I didn’t see this one coming and it caught me off-guard if I’m being totally honest. Just when I thought I was getting my head around containers and K8s, Google show up to the party and drop this on us. Baked into the existing Cloud Functions service, serverless containers will let you package a function within a container, to better support custom runtimes, binaries and frameworks.

This means you’ll no longer be restricted to a language, or a version, nor any type of environment. It’s a completely different way of looking at serverless computing. In fact, this particular announcement led the Servian team to having a lengthly discussion over dinner one evening about the state of play and how things will pan out off the back of this news. It was actually the best dinner we had all week :-)

Having serverless containers opens up so many new doors and opportunities, but it also raises a few questions for me at the same time, like: what will the spin up times be like? where does this now leave Cloud Functions? what will the costing model look like?!

Note: it’s in alpha sign up, so here’s the form to request it.

Presenting at the Community Lounge

This was great fun — even if I was a little hungover!

I was honoured to have been given a small speaking slot at the Community Lounge, where I delivered a lightening talk on how to run a successful GCP user group in the community. I ran through our top 5 tips that work for us:

  1. Build a diverse team around you, and rotate the duties each event.
  2. Connect with local Googlers, Partners, and other user groups.
  3. Provide members with an easy way of giving feedback — and listen to it!
  4. Having a consistent venue and date each month helps with attendance.
  5. RSVP numbers lie — only 40–50% actually show on the night!

In general I received good feedback for my talk, even though I was mildly hungover from the night before. That said, one person did come up to me afterwards and say he disagreed with 4 out of 5 tips. High five!

He obviously missed the intro where I said “these tips won’t necessarily work for everyone!” ;-)