I like to keep my finger on the pulse of technology. And for me, this means more than just reading about tech. I often don’t quite grasp the power of new tech, the possible use cases, the potential impact, until I try it. Here are some things I’ve been playing with recently.
Flow Blockchain and Cadence
You may have heard of NBA Top Shot, an NFT-driven marketplace of basketball video highlights. It runs on the Flow blockchain, developed by Dapper Labs in partnership with the NBA. Dapper built Flow to address the performance issues it experienced with its CryptoKitties game running on the Ethereum blockchain back in 2017. The concept behind Flow is the use of the consensus model only for the tasks requiring consensus, enabling the other computational tasks to be done more efficiently.
I wanted to try to build something on Flow – a smart contract, an NFT, a fungible token, anything really. The great thing is Flow has a tutorial which lets you do all of these things.
The easiest way to do this is with Cadence, the native programming language for building Flow apps. One of the cool things about Cadence is its resource-oriented programming model that ensures a resource, which represents a scarce digital asset, can only exist in one place at a time. It can’t be copied and can’t be accidentally lost or destroyed. For example, rather than typical programming language assignments that copy values or objects, in Cadence you do a move instead:
It takes a bit of practice to get used to the model of creating and executing smart contracts and transactions, and moving and storing assets. But I like it a lot and want to find some use cases and build something a bit more complex.
SwiftUI
SwiftUI is Apple’s new UX framework for iOS, macOS, tvOS, and watchOS apps. SwiftUI and Xcode provide an incredibly powerful platform for building interactive apps.
SwiftUI is a radical departure from its predecessors UIKit and AppKit, in that it is a declarative programming model. You use SwiftUI to describe what your UI should look like, and Swift handles the how. The basic UI component, a view, is often said to be a function of its state. Change the state and the view is automatically updated. This is similar to declarative web frameworks like React.
Writing your first iOS app with SwiftUI literally involves writing zero lines of code since a “Hello world” app is the starting template for building any app. But doing things beyond that start getting complicated, and it takes a while to understand SwiftUI conventions and learn the nuances of Swift. And certain things can’t be done in SwiftUI, requiring use of UIKit and AppKit and view delegates and other aspects of Apple’s former UI model.
I don’t know how far I would have gotten if it wasn’t for the dozens of tutorials on Paul Hudson’s Hacking with Swift site. Paul’s videos are among the best tutorials I’ve seen on any technical topic, period. In a relatively short amount of time I was able to build an app with:
- An Apple map with location annotations and location tracking
- Ability for users to upload images and take and store photos
- Storing submitted photos in a popular photo-sharing website
- Storing and retrieving information to a no-SQL database
- And of course the basics of displaying information in different layouts and showing and validating forms.
I will definitely come back to SwiftUI. It’s truly a delight to work with.
Amazon AWS
Amazon’s suite of cloud services is really impressive. Many of the services by themselves, from cloud compute to container management, analytics to machine learning, are among the best in their product classes. But the real value of AWS comes when you stitch the services together.
There is so much to AWS that it’s hard to decide where to start. So I just dove right in to something that looked interesting, figuring the different pieces and context will become clearer as I go. One of the things I wanted to learn about was how to consume and analyze large volumes of data. I found a one-hour lab on AWS called Analyze Big Data with Hadoop. This lab required purchase of 1 credit which amounts to around $1.
In less than an hour, you use the Amazon EMR set of tools to deploy a full Hadoop cluster and run some HiveQL to process sample server log files. HiveQL is a SQL-like query language that lets you avoid the complexities of lower level programming for Hadoop. The files are read from an Amazon S3 bucket and the resulting analysis is stored in files in S3. Although a rather simple example, you can see how this could be scaled to process huge datasets. The one-hour intro barely scratches the surface on Hadoop, Hive and Amazon EMR. If anything, it gives you a glimpse of the power of the ecosystem and leaves you wanting to learn more.
I’ve gone through a few other AWS tutorials – how to set up a virtual private cloud (VPC) and how to configure Linux virtual web servers among others. So much to play with, so little time…
Edge Computing
On to the next computing paradigm – edge computing! The edge concept is to do some or much of the computing near the data sources. Data sources are usually tied to the physical world, for example a temperature sensor. Rather than stream all the data from the sensor to the cloud and process it there, why not do some of the processing right in the sensor or in a local server near the sensor. This can help with the typical network issues of cloud computing (latency, reliability, etc.) and better utilize the processing power available with edge devices.
The downside of course is software management. Who wants to configure and maintain the right software stacks on all the devices that are part of your computing ecosystem? This is where containers come in. A container is a complete package of an application, its configuration and dependencies, all wrapped up and easily deployable. Container management tools have been around for years but sadly I haven’t had a chance to use any of them. Until now!
If you’ve heard of containers, you’ve probably heard of Docker. Docker gives you a robust set of tools for packaging applications as portable container images. It makes it super easy for teams to build development and staging environments, and ultimately deploy their apps to production. What’s really great is the ability to download container images of common platforms from Docker Hub. Want to set up an ubuntu instance on your computer in seconds? Or a database like MySQL? And you want to bundle these together as part of your application? All easily done with Docker (once you know what you are doing). There is an excellent tutorial on docker.com that has you containerize an app, upload the image to Docker Hub, and use Docker Compose to create a multi-container app. Very easy to follow. And I must say, it seems much easier to start with a container image rather than downloading operating systems and databases and programming frameworks and getting them all set up and configured. I wish I had tried this earlier.
The tutorial is great, but I wanted to try something on my own – getting a version of Redis running on a Linux server and saving some data to it using the Redis command line interface. I grabbed the latest release of Redis from the Docker Hub and it installed in a few seconds, OS and all. After a bit of experimenting, I figured out that I needed to create a docker network and run two containers in that network, the Redis server and Redis CLI. Just three Docker commands from my PowerShell is all it took. Then I could interact with the CLI to store and retrieve data from Redis:
Amazing! But enough of that for now.
Another big name in containers is Kubernetes, a container orchestration tool. This is where I am heading next…