AWS Summit Online Australia & New Zealand: Recap
I have had the pleasure of being a part of the team that has been representing Idea 11 at the AWS Summit Sydney for the last couple of years and while the team was getting ready for Summit 2020, the world changed dramatically for everyone and unfortunately resulted in the cancellation of all mass events, including the AWS Summit.
A month later, AWS announced that the AWS Summit will be streamed online as a virtual conference. I was extremely happy to see the invitation in my inbox and soon completed the online registration process.
Now, if you haven’t attended an AWS Summit before, let me get you up to speed. An AWS Summit is a free event designed to bring together the cloud community to learn and connect with AWS and its partners. AWS Summit Sydney is one of the largest events of the year in Australia.
The AWS Summit Online followed the same format as most technology events, keynote followed by breakout sessions.
This year’s keynote was presented by Dr Werner Vogels, VP & CTO of Amazon. The major focus of his talk was around the significance of architecting technology solutions and building a solid foundation using tools like the AWS Well-Architected Framework
Here are my top 6 takeaways from this keynote:
- AWS Builders Library - This contains some extremely detailed papers on how Amazon builds and operates software. This library provides free access to a collection of articles written by the company’s senior technical leaders on how they design, build, run and scale the hardware and software systems underlying Amazon.com, Amazon Web Services and Amazon’s other businesses.
- AWS Managed Services like Lambda, Aurora, DynamoDB, Connect and Workspaces allow you to scale and build faster.
- AWS allows you to pivot and try new ideas quickly (way faster than you ever could on-premises and for way cheaper too). This also allows you to innovate and respond quickly to customer demand.
- Dr Vogels’ Now Go Build web series on start-ups all over the world in some really interesting locations.
- AWS Training & Certification – There is a huge amount of free training available here, that is built by the experts at AWS.
- This is my architecture – A video series by AWS on innovative architectures. It is a great platform where companies talk about how they have built their cloud architectures using AWS solutions.
Now came the time for my first breakout sessions. When choosing my breakout sessions, I like to pick a few topics that I have absolutely no idea about (we are here to learn something after all) and pick some subjects I’m personally passionate about to see if there’s anything new I need to know (because there always is).
Building Serverless Applications with AWS Amplify
So, my first session was by Derek Bingham on 'Building serverless applications with AWS Amplify'. To give some background if you do not know me, I am not a developer, but I like to know how things work. So, this session was certainly more towards the “I have no idea about” side and it turned out to be perfect.
AWS Amplify is a toolset that allows you to build serverless applications quicker. It is a development platform that allows you to build mobile and web applications. What’s really cool about AWS Amplify is how it has been built to abstract the developer from having to worry about the underlying infrastructure. You get to build the application; AWS looks after the infrastructure, helping with things such as data storage, machine learning, and much more.
With the release of the new native Amplify SDK’s, it’s making things even more simple by abstracting the underlying infrastructure even further.
Now I know this may feel like you will be limited in terms of what you can do, after all as Derek puts it “AWS Amplify is an opinionated framework with AWS making assumptions to make things easier for you”. However, AWS has added a feature called the “Escape Hatch” allowing you to now do what you need to do.
Mastering Your Data Journey One Step at a Time
My second session was by Rada Stanic on ‘Mastering your data journey one step at a time’. Rada has been helping customers modernise their data strategy for quite some time now and strongly believed that the data we have today no longer fits in the traditional data warehouse. Enterprises now want to be able to have a data platform that will live for 15 years and scale.
The focus needs to be on a data framework, not the technology, no longer is it just BI’s and BA’s accessing the data, its now data scientists and machine learning specialists. Therefore, companies need a platform to support all the use cases. This is where a data lake comes in. For anyone wondering what a data lake is, here’s AWS’s definition “A centralised repository that enables you to secure, discover, share and analyse structured and unstructured data at any scale”
What I really liked about Rada’s talk was how she explained the process of incrementally moving to your new data lake and how AWS Lake Formation can help automate a lot of the heavy lifting.
Cloud Security for Everyone: Multi-Account Strategy
My next session was 'Cloud security for everyone: Multi-account strategy' by Byron Pogson. Byron had a great story to tell.
He talked about his start-up - ‘Salt as a Service’. This small team started out with a single AWS account housing all their development, test, and production. But with their new product ‘Pepper as a Service’, they needed to improve. They needed isolation of their security and audit functions. They needed identical environments so they could do full-scale testing without affecting production. They also needed to provide better cost allocation to identify what each environment was costing them.
By now, you must have realised that this is not a real start-up (at least I could not find it on google) but this is a common story Byron tells his customers.
Byron also highlighted the importance of services like AWS Control Tower. AWS Control Tower allows you to quickly set up a multi-account AWS architecture complete with governance and guardrails.
The best thing about Control Tower being a managed service from AWS is that when AWS releases new best practices or controls, you can just update your Control Tower and you have them too.
Now, this may sound daunting in an existing environment, but it is possible to roll out AWS Control Tower in an existing organisation and is something I have enjoyed talking about since its release in the Sydney region in March.
Advanced VPC Connectivity Patterns
For the networking guys reading this, Brett Looney had a great talk on Advanced VPC connectivity patterns. His talk covered some great topics such as:
- Connecting to AWS with AWS Direct Connect, including why using Direct Connect Gateway can help simplify multi-VPC and multi-region connectivity.
- When to and when not to use Direct Connect Public VIFs?
- VPC Peering (and its limitations) and how AWS Transit Gateway can help you overcome these as well as simplifying your network architecture.
- How you can use AWS PrivateLink to provide services into other AWS accounts (like your customers) without going over the internet?
AWS networking gets updates and upgrades all the time, so this is a great talk to watch if you need to get up to date or you have a new challenge you need networking answers for.
How to Drive Economic Value through Cost Management and Optimisation
Cost optimisation is a topic I’ve always had a bit of a passion for, so if there is ever a talk on it I’m always looking for new perspectives or guidance and boy did Nathan Besh’s ‘How to drive economic value through cost management and optimisation’ deliver!
A few things in this talk really stuck with me:
Cost optimisation is not just about making my bill cheaper the next month. Cost optimisation is the ability to run systems to deliver business value at the lowest price point.
Quick wins are a measure of your lack of capability. Sure, it's great to be able to reduce $10k a month off your bill, but why was it there in the first place? How long has it been there?
Do you have a security team/function? Do you have an operations team/function? Do you have a cost optimisation team/function?
Stop talking about billing, look at the value and efficiency of what you are doing. You came to the cloud to do different things and more things that you could before. Does it matter if your bill goes up by 10% if your customers served went up 400%?
If you expect people to build cost-effective systems, are they driven to do that? If they build a system that costs too much money, is there actually any impact on them?
Do not pay 3 times for a mistake:
- Provision it incorrectly, every hour you are paying for it.
- You then must find the waste either paying for a tool or people.
- Then you must pay to fix it.
Nathan also touched on the new Savings Plans and how to properly take advantage of them.
Operations for Serverless
At Idea 11, I have had a chance to wear a few hats and a few years ago, I used to run our TechOps managed services team. So, learning new ways to help operational teams always interests me. Chandra Allaka’s talk on ‘Operations for serverless’ certainly showed me some tricks.
Idea 11 does a lot of serverless with our development practice, but this always raises the question, how do you monitor and troubleshoot an application when it can be made up of 100s of microservices.
Chandra gave some great advice on key operational challenges, covering:
- Dependency management and how you can create a service map to identify how everything hangs together. Dependency management is key to issue resolution and change control.
- Everything breaks, but how do you identify where it is breaking? You need to introduce observability into your application with things like log correlation and services like AWS X-Ray
- Change Management – You need to classify your changes, try, and do small and frequent changes and reduce risks through versioning and canary deployment features.
Building Resilient Applications Using Chaos Engineering on AWS
For my last talk, Adrian Hornsby on 'Building Resilient Applications using chaos engineering on AWS'.
Ask yourself these questions:
- Do you train for outages?
- Have you trained how to understand outages?
- How do you debug outages?
- How do you do a post-mortem?
Now how do you train for outages? Well, you need to simulate them. In comes ‘Chaos Engineering’.
As Adrian puts it - “Chaos engineering is NOT about breaking things randomly without a purpose, chaos engineering is about breaking things in a controlled environment and through well-planned experiments in order to build confidence in your application to withstand turbulent conditions.”
Chaos engineering is a scientific method and you must believe that your environment is already resilient before you start. Chaos engineering is about discovering the unknown, and this is the real difference between testing and chaos engineering.
Adrian goes on to talk about how to implement chaos engineering at your workplace, so if you are ready for the next step in building resilient systems, give this talk a watch.
Even though I know everyone was excited for another few days in Sydney (and getting one of Idea 11’s famous plush dinosaurs) it is great to see that we can adapt to new circumstances and still have an event like AWS Summit Online. I really enjoyed getting to watch these presentations and look forward to seeing what is next.
The AWS Summit Online content is now all available on-demand, so be sure to check it out.