Homegrown Latency Cloud Conference 2020

In this year of the COVID-19 pandemic, many in-person conferences were cancelled or reworked into online events. Our home town of Perth went through a period of lockdown earlier in the year, and implemented strict border controls that contained the spread. Easing restrictions meant the Latency Cloud-native conference could be held as a live event. Aside from its Homegrown theme—featuring local speakers—the conference felt just like previous years.

Adapptor sent another contingent to stay sharp on new technologies, see what our peers have been working on, and catch up with local industry happenings. Below, our developers share some highlights from Day 2 of Latency 2020.

Jonathan Tennant

After preliminaries, we were treated to a poignant and wonderfully personal Welcome to Country message from Betty Garlett. The following keynote by Dr Liz Dallimore from the West Australian Data Science Innovation Hub introduced the Design Thinking Process. Her talk highlighted its first two phases—Empathise and Define—and demonstrated how 7 principles can help isolate issues at each stage of a user's journey that offer opportunities for innovation.

Next, Bruce Dominguez of VGW spoke on the topic of incident response. After striking fear into our hearts about the many ways a poor incident response can leave an unprepared tech caught like a deer in the headlights, Bruce expounded the virtues of preparedness and learning from mistakes. He tackled the problem of expecting-the-unexpected using tools from a surprising quarter: tabletop role-playing games. He described a method of practicing for real world situations using a gamified simulation of how a team would handle an imagined scenario. The team uses real world actions and abilities captured as playing cards to respond to events and problems thrown at them by the “Dungeon Master” (or rather, “Incident Master”). An innovative approach to solving a difficult problem.

Noel Waghorn

Shermayne Lee presented a tracking app for her (adorable) dog, Puffy the Pomeranian. Starting with a cheap, small LoRaWAN—low power, long range, wide area network—tracker attached to the tiny dog’s collar, she registered the device on a global LoRaWAN network called The Things Network. The network allows information to be sent to and from the device, in this case geo-coordinate updates. Information on the network is queried by AWS API Gateway/Lambda and then stored in DynamoDB. The device’s latest location is retrieved by AppSync and presented in a web browser using React. The takeaway: Shermayne believes IoT projects are much easier to implement than they might at first seem—that and her dog deserves fame and fortune. With a bit of research and planning, many pieces of the puzzle already exist, just waiting to be used.

Dan Venkitachalam

A highlight for me was ex-colleague Brian Foody's talk on the Compounding Returns of Serverless. If you follow Brian's profile on LinkedIn, you'll know he's a keen student of investment strategies. His talk focused on the magic of compounding, where seemingly small initial investments yield exponentially higher returns over a long time period. On the flipside, fees that appear relatively small can accrue over a long period to have a significant impact on your overall yield.

This is a well-known principle in personal finance, but how does it relate to software development? Brian made the analogy of a software project as a long term investment. Implementing supporting infrastructure such as automated testing and deployment can be seen as upfront investments that won't move the needle in feature delivery. But through compounding effects, these early decisions pay off to free teams from the growing overheads of maintenance that cut into development time.

Similarly, taking on technical debt through development shortcuts can be seen as a small overhead cost initially, but one that incurs a compounding loss over the product lifecycle. Brian argues that utilizing integrated stacks such as Serverless reduces the costs and compromises that are typically made early in development, and pays dividends as their effects compound over time.

Kurt Blackwell

For “How to Build and Manage your IoT Project from PoC to Production”, Nathan Glover presented a product from Planet Toast: “The world’s most innovative Internet Connected Toaster company”. He took us through the series of choices required to turn an ordinary toaster into an internet-connected, AI-powered device, and discussed popular hardware platforms, connectivity solutions, and integrations with the Cloud. To control and manage these devices he used AWS Greengrass, and introduced Certificate Vending Machine as a way to secure them. This presentation was similar to his AWS Community Day 2020 talk, Top 5 Tips for Scaling your IoT Projects, but served as a better introduction for those less familiar with IoT devices.

Jason Crowley

Luke Mundy from VGW gave an interesting talk on why he built an AWS Step Function. To debug issues, VGW often required data from production, but getting the data was risky. It involved manual manipulation of their production environment, snapshot restores, data sanitisation, and even database deletion. To automate this process and make it safer, an AWS Step Function was created, and he described the complexities of configuring its security requirements. Being unfamiliar with Step Functions, I was interested to hear a complex use case, and it piqued my interest to research them further.

Simon Bradley

In “Build Components Not Apps”, Alexander Karan urged us to put our efforts into first building a component library, before using elements from this library to compose our apps. He touted the benefits of this approach, compared to jumping straight in to coding your app, including: reduced code duplication, dramatically reduced development time for subsequent apps, UI consistency, and ease of theming entire apps. Components can be UI elements, but they can also encapsulate discrete pieces of logic, such as networking or database access. Suggested tools to assist with this approach include Bit, an online component repository with features such as dependency management and automatic versioning, and Storybook, a rapid UI development tool for React.

With COVID-19 travel constraints in place, Latency featured an all-local lineup, as suggested by its “Homegrown” theme. Despite this, the quality of talks remained high, a reflection of the depth of knowledge in the Perth Cloud-native industry, and the democratisation of technology that global cloud service providers have enabled.

The range of topics falling under the Cloud-native umbrella was wider than ever. As app developers, we found the backend talks most relevant to our day-to-day work. The challenges and opportunities of Web and IoT development have many parallels with mobile app development, and it would be nice to see more time given to mobile apps. This convergence can be seen in technology stacks such as Amplify and Firebase, which continue to extend their app capabilities.

We’re grateful to Mechanical Rock for organising this unique event, and look forward to the next iteration.