This series of blog posts intends to share real life experience with an enterprise lift & shift project from on-premise data centers into Google Cloud Platform (from now on referred to as GCP).
Don’t read this
- If you want to know what Cloud is: there’s always this.
- If you’re comfortable with your on-premise and not looking for innovation: still using floppy disks? Don’t worry, there’s always been people like you (please, do read the entire article. It’s great).
- If you want to learn the basics of GCP: there’s other places for that.
- If you want to know which Cloud to choose: I’m not a salesperson. Just a GCP fan.
Do read this series…. if you want to be taken onto an exciting journey of real-life experiences moving entire workloads into GCP. I’ll start with a bit of background to get you all on the same page and hopefully ensure everyone clearly understands where we’re coming from.
A Cloud the world wasn’t ready for
Google has been running their ‘cloud’ since the beginning. After offering SaaS solutions like Gmail, Maps, Youtube to the public, Google started offering the first parts of their Enterprise Cloud with App Engine (2008) and BigQuery (2010), both fully managed services. With App Engine, a PaaS, you just push your Java or Python code for your webapps, and it will handle the rest. Yes, all of it. Want to do (big) data analytics with BigQuery? Just upload your enterprise data warehouse data and start querying it. Petabytes of it. In minutes. Yes, that’s it.
Back then and still now, businesses are not familiar with that level of ease. I even spoke to a prospect last week who couldn’t wrap their head around serverless solutions like BigQuery. “Will we get notified of downtime for OS patches and maintenance windows?”.Yes, this is 2018 I’m talking about. Clearly some people are having a hard time letting go.
For most businesses, GCP’s level of abstraction was too high, it was too simple to use, the tedious details were too automated and there was no way of just switching over to it. It was Google’s biggest mistake thinking the enterprise world would be ready for these advanced, but then again simple, technologies. Most of them are still jamming their AS400 terminals. And let’s face it, they love it.
Google’s Tinder-matches
Google’s profile of a Cloud provider was too fancy. People were scared of it. “It’s good for a startup to start from scratch, like snapchat, but this is real-life. We have tons of servers and integrated applications running our production environment”. Valid point if you ask me. Not having a clear entry point into GCP prevented a lot of companies to even look at it. How would you make single public-facing App Engine applications fit in with your existing Enterprise Architecture?
Well, that changed. To make sure that they weren’t swiped to the left by 90% of the enterprise market, Google had to make some changes. They introduced lower level solutions like Compute Engine (2012), Cloud VPC with VPN capabilities, and shared solutions in between PaaS and these IaaS, like Kubernetes (2014), to be a lot more attractive to a broader range of companies looking at Cloud solutions.
Now companies can simply migrate some VM’s into Compute Engine and start exploring from there. Maybe move the MySQL into Cloud SQL as a first touch on managed services? A lot less scary with these baby steps.
Google also took a different marketing approach. Rather than waving with all the no-ops, expecting companies to drop everything they were doing and just replatform, (ever tried moving your JEE stack from Websphere to App Engine?), Google’s message got a lot more acceptable. “Meet you where you are and take you further”. Let’s say this upgraded profile got swiped “to the right” a lot more than before.
Taking the red pill
The last few years, we have seen an amazing uptake in the adoption of Google Cloud. I’m not about all those different metrics, just have a look at their current customers (hey, is that another Fourcast customer on there?) Coca-Cola, Best Buy, Philips. Wouldn’t call them “greenfield startups”. These testimonials are of vital importance, showing large multinationals also have a use case for GCP. At first, they might only do it for cost reduction, but the real gain is in the long term. Infrastructure is just that, infrastructure. This is the part where Google met you where you were. Now you need to let them take you further.
Artificial Intelligence, Big Data, Machine Learning, Neural networks. These are the real hidden gems of the Cloud. Processing million of customer transactions, matching them up with products, to give every single customer a hyper-personalised experience, in seconds? Distinguishing clouds and snowy mountaintops on satellite images? Jup, that’s all possible now. Just lookup TensorFlow, and how Google can do the heavy lifting for you with AutoML (2018).
Do you think Neo would ever go back to being Thomas A. Anderson?
Now we’re all taking at the same level, you’re ready to start exploring the ins and outs of an actual project, migrating an entire datacenter from on-prem into GCP.
Next up: 2 – Looking through the fog
