Cloud Hosting Providers: what you need to know.

22 April 2024 by Catalyst

The cloud infrastructure market can be confusing and having to make decisions about things you don’t fully understand is nothing but frustrating.

This blog is not about selling any of the big cloud providers, as many of us only have a small number of options when considering the required feature set, locality (data sovereignty laws) and capability of an internal team or contracted provider. Rather, this blog aims to provide some pointers – things to consider – when making high level decisions such as moving to cloud and choosing (or changing) a cloud provider.

We sat down with Andrew Boag, Catalyst IT Australia’s Managing Director, to pick his brain on some of the common questions about cloud providers. Having been in the industry for over 20 years and managing multiple teams and clients in different regions, we hope that Andrew can help clarify some things.

Q. What difference does a cloud provider make anyway?

At the very least, organisations, including Catalyst, have tired of the responsibilities of managing physical hardware in a data centre.

Catalyst had several racks of physical machines until 2021 and yes, there were some considerable cost efficiencies for some workloads – cloud isn’t always cheaper when running on your own “tin” servers. But the annoyance when pieces of this infrastructure broke, which is an inevitable eventuality, was part of the motivation to move pretty much everything into cloud.

While companies and governments have not turned off all of their data centres yet, more and more of the compute and storage that drive the world we live in, and the smartphones that we depend upon, are housed on one cloud platform or another.

Data centres are not dead, but the trajectory of all the data and all these services is unambiguously heading towards cloud providers.

Q. How similar or different are cloud providers?

Comparing cloud providers will lead you to an epiphanic duality where cloud platforms can be both “quite similar” and “very different” at the same time. Very Quantum theory.

Of course cloud sales people will tell you exactly what you want to hear about the advantages of one platform, or the ease of migration from your current platform into theirs. This can all be discounted.

Different cloud providers are all similar in that they will facilitate you standing up a Linux or a Windows virtual machine i.e. compute, to talk to a relational database and some objects or file service i.e. storage. You will be able to run pretty much any application that you used to run yourself in-house. That is a given.

However, the differences are very real as soon as you take a step deeper into all the powerful services and toolsets below the surface that cloud providers offer.

This can be anything from the cache services supported to the logic around high availability database design.

There are also, and always, some as-a-service offerings that will completely differ across providers.

One might have a far better image recognition service than the other, for example, and they will be implemented completely differently.

How many times have we sat through a meeting or a sales pitch reviewing the “cloud options” for our organisation. Should it be Azure, AWS, Google Cloud, Digital Ocean or a local OpenStack or Kubernetes provider for our cloud infrastructure and tools?

The cloud infrastructure market is very top heavy, and no one ever got fired for choosing Microsoft or AWS.

Q. What about Multi-Cloud or Cloud Agnosticism?

This was a big topic a few years ago. With a lot of cloud conferences talking about “hybrid clouds” and the flexibility that comes from being able to consume services from various providers.

This was a lot of the appeal of Kubernetes in the beginning – in that all of the cloud providers supported it.

However, from Catalyst’s experience there are considerable challenges to having mission critical web workloads running across disparate and different cloud platforms. There are always numerous things that need to get re-implemented, with considerable time and cost implications.

A good example are the tools that allow you to do availability monitoring of your cloud infrastructure. There may be large differences in the mechanics of querying state between Azure and AWS for example, which means your team will waste resources, doubling up on building and maintenance tasks.

There is also the very practical consideration of internal team expertise. We were recently approached with an opportunity to implement a Moodle solution on a new cloud platform (that we don’t use). The reality is that we would be taking someone who is a seasoned expert with one cloud platform and taking them back to the newbie level while we figure out how it all works.

Cloud platforms are so big that it’s extremely unlikely that you’ll find someone who is an expert in more than one. This has real implications for your ability to support mission-critical sites 24 hours a day, 7 days a week. Do you really need two separate teams maintaining your sites?

Q. What to do when the incentives for change are too tempting?

We have also seriously considered “jumping ship” from one provider to another. Including discussing this at a high level and securing sizeable discounts and migration credits. It is, however, a massive change and with the initial round of planning, we realised that we would actually be running two cloud platforms for a number of years while we fully migrated according to some contractual constraints. This is not what we wanted even with considerable, sweet sounding cash incentives.

Q. So how do you decide – AWS, Azure or Google?

Very good question and you should listen to their respective pitches!

You should sit down and map out some back-of-the-envelope migration projects into said platform. Then, triple those numbers!

You must listen carefully to what your operational and technical team say about their preferences – it really matters.

My biggest advice is: reach out to the [cloud migration] experts in the consulting space. There are a number of quality operators out there that provide cost effective “heavy lifting” – take advantage of that, especially if you are only in the beginning of your cloud journey.

Q. What has the Catalyst team learnt from 15+ years of Moodle on AWS?

Catalyst has been running Enterprise Cloud Moodle LMS instances on AWS since approx 2008. Actually, earlier if you count some bespoke consulting engagements in the US.

There are some super affordances for running an LMS on dynamic cloud infrastructure as, no surprise, there are quite predictable and meaningful differences in application base load at different times of the day.

For a large University, for example, the local night time is considerably less busy than during business hours. This means that we can optimise our spend and cloud resource usage based on the patterns that we observe and track with reporting, monitoring and analysis.

We are constantly and consistently implementing improvements to deliver performance, load capacity and optimise AWS spend footprint. Keeping an eye on client usage patterns and responding with varying operational strategies has helped us maximise performance while managing budget.

Over the years, we have experimented with other cloud platforms – Azure and Google Cloud. While these platforms are certainly capable, we have estimated considerable “cost of change” if we were to move our workloads from one to another.

So what have we learnt – One, there is very low appetite to be across multiple cloud platforms and two, it’s about quality over quantity. We have built and continue to progress our expertise in the AWS platform, mastering all the tricks of the trade and passing on our lessons – time and cost savings – to our clients.

Like you said, this blog is not meant to be selling any particular cloud provider, but to draw attention to the following facts:

  • the shift to cloud is inevitable for most organisations
  • it does take time to learn your way around cloud infrastructure before you can make it work for you in the most efficient way
  • moving from one provider to another comes with many costs (visible and hidden)
  • taking advantage of a cloud service provider with mature infrastructure, who have been in business for some time and are trusted by organisations similar to yours is really worth it

We hope this helps answer some of your questions when it comes to making decisions around cloud hosting and cloud providers.

Check out the following blogs we’ve written previously on Cloud and AWS more specifically:

Understanding managed services in the cloud

Cloud computing – risk and mitigation.

Data protection with cloud backup as a service.

Protect your enterprise Moodle with 24/7 support from Catalyst.

What is cloud automation and orchestration?

Are you paying enough attention to your AWS spend? A few tips on how you could save thousands.