This week Pagely announced NorthStack: a managed WordPress hosting product that isn’t just new: it’s touching on revolutionary.

Performance Foundry has been, and continues to be part of the NorthStack alpha program, and we’re looking forward to being early adopters when it’s production-ready. We can see great applications for many of our clients across e-commerce, membership sites, and publishers.

What is Northstack in a nutshell?

It’s pitched as:

A new managed hosting service featuring elastic autoscaling, serverless infrastructure, and simple pay as you go billing.

So what’s so special about NorthStack? There are a couple of important ways this varies from standard hosting, both technically and in how its billing is set up.

Serverless WordPress

Serverless is a big topic, and one we’ll be writing more of in the future. To sum up, we can say that serverless computing is about only using the resources needed for each individual computing request.

Instead of having an application sitting on server ready to serve data, there’s a configuration file that automatically creates micro-instances of servers and runs just what’s needed to serve the data.

It’s the equivalent of using Uber instead of owning a car. If Uber vehicles and drivers only existed on demand, and processed everything in microseconds.

Serverless WordPress is hard! WordPress itself was designed and built in the years before serverless was a valid idea. The hardware and systems definitely didn’t exist!

Up until now, I’ve never seen serverless WordPress done properly. Many solutions look to create static versions of the WordPress site and just serve that, leading to multiple issues with interactions – like making sales or leaving comments. NorthStack still has its limits, but it is moving beyond that.

Elastic Autoscaling WordPress

A lot of smart people have combined both Docker and some kind of infrastructure to scale the CPU needs of WordPress. That still leaves us with always-on resources: base CPU, image storage and databases are still running in some kind of dedicated/VPS/containerised/virtualised environment. Then there’s gateway access, security, performance monitoring… all that background stuff you don’t think of. These resources still need to be manually managed.

But truly elastic autoscaling has no limits. Even the best scaling hosting I’ve seen to date struggles with the database element. It’s really hard to scale WordPress databases. You can go horizontally, adding more clones, but WordPress’s architecture doesn’t lend itself to that. You can shard it, putting some tables on different databases: that works for multisite, but not so well for large single instances. Then you can scale vertically to more resources – but there’s always risk in migrating, upsizing or downsizing database instances.

In the case of NorthStack and serverless WordPress, this is all handled by Amazon Web Services (AWS)’s auto-scaling technology. Their brand new Aurora scaling database services are the key to unlocking the potential of WordPress and serverless.

Pay as you go billing

NorthStack’s turning the managed WordPress hosting proposition on its head. Instead of dedicating hardware resources and billing for them, serverless allows everything to go ‘on the meter’. Pay for what you use.

And that’s going to go for NorthStack support too: support access will also be handled on a pay as you go basis, with different tiers and access points. It’s the only thing that makes sense.

For Performance Foundry, this will introduce a few interesting discussions around how we bill too! Our WordPress retainers are going to be a big part of that. For sure, NorthStack isn’t going to be the best solution for all our clients, but there are really strong business cases for moving some of our clients over as soon as we’re happy with the project being production-ready and stable.

What’s Performance Foundry’s involvement?

We were pleased to be invited to the NorthStack alpha release meetings, and were one of the first agencies to have access to the code. NorthStack’s philosophy matches with ours: high quality infrastructure at a price that meets business objectives.

At the moment, I (Managing Director, Craig Martin) am experimenting with the early code base and following development. Our goal, as the code becomes more steady, is to train all our development team on NorthStack before its production release and partner with our clients to use this technology when it’s a good fit for their projects.

The future is both autoscaling and serverless. We’re excited.