AWS in 2019 – Year in Review § Predictions for 2020

January 7, 2020 Author: Rado Minchev Category: AWS

As 2019 ended, I decided to review AWS’s activity this past year.

“Reviewing” turned into hours of reading and watching videos.

Even for someone working with AWS daily, it’s impossible to keep up with everything.

I’ve compiled a list of news, courses and services/features that I found interesting.

This list is by no means exhaustive. If you have a billion hours available, check out AWS’s official website.

Let’s get started.

Savings Plans

I feel like this new pricing model got lost in the array of news AWS puts out daily.

And that’s a bad thing.

Savings Plans are a game-changer for pricing. They might even spell the end for Reserved Instances.

Savings Plans provide a few vital advantages:

You commit to a level of compute usage per hour instead of purchasing Reserved Instances for a given type. This saves time and effort for everyone and especially for new users.
You pay a lower (Savings Plans) price up to the amount you committed to. All usage above that price is billed at the on-demand pricing level.
You can apply Savings Plans to both EC2 Instances and Fargate.

I’ll get into the details of AWS pricing in a separate article.

If you want a more detailed explanation, check out the AWS website.

New AWS Courses and Certifications

It’s worth noting that you can finally take the Cloud Practitioner Exam (the entry-level certification) from your home/office. The process is still far from perfect, but AWS and Pearson Vue are taking steps in the right direction.

Machine Learning

AWS (and everyone else) is betting big on ML. They’ve added a bunch of new services (as you’ll see later) and a course on building and training a CV model using Apache MXNet and GluonCV.

Finally, there’s a new AWS specialty – Certified Machine Leaning.

Alexa

Alexa is Amazon’s AI assistant that’s used in the Amazon Echo/Echo Dot. AWS is trying to fill the voice skill gap by releasing an Alexa for Developers Course that explains concepts like Voice User Interface and the usage of the Alexa Skills Kit.

Similar to ML, they’ve also added an Alexa specialty – AWS Certified Alexa Skill Builder.

AWS Fundamentals & Others

AWS has also introduced courses and workshops related to their core services. The learning curve is still pretty steep for newcomers, but releasing more free content definitely helps.

AWS Developer Series on edX – This is by far the most comprehensive series of courses for developers released by AWS in 2019. It’s broken down into 3 modules – Building/Deploying/Optimizing on AWS, which you can take in any order.

The rest are honestly pretty self-explanatory, so I won’t ramble on about them.

New Services/Features

The total number of AWS services is now approaching 200. For comparison, in mid-2017, that number was less than 100.

Good luck to anyone starting out.

Obviously, I can’t cover each new service or feature here. Instead, I tried to summarize and combine together the most important ones by topic.

Deep Learning, Machine Learning and Artificial Intelligence

As I said, AWS is going big on DL, ML and all other acronyms related to Artificial Intelligence.

AI Speech Analytics – this solution provides you with a pre-trained AI that transcribes, translates, and analyzes customer interactions in Amazon Connect.
SageMaker Studio is AWS’s attempt to combine everything ML devs and data experts need in one place. It includes things like a fully managed ML service and common algorithms while providing easy deployment into a scalable environment.
Amazon Augmented AI is built for systems that require human reviews of ML predictions. The main benefit here is that AWS removes a lot of the heavy lifting that goes into creating a human review workflow.
Fraud Detection With Machine Learning is the hot new trend in financial services. AWS is jumping on the bandwagon with this solution that automates the detection of fraudulent activity and also contains an example dataset of credit card transactions.
Deep Learning Containers – if there’s a name that tries to capitalize on hype more than this one, I haven’t heard it yet. These are Docker images with pre-installed DL frameworks.
New EC2 ML Instances – getting started with ML on AWS was a lot tougher in January of 2019 than in December of 2019. Along with Compute Optimizer and SageMaker Studio AWS has made things a lot easier this past year. These new EC2 Instances (named Inf1) that feature the AWS Inferentia chips are the finishing touches.

Cost Optimization, Backups and Other Best Practices

Cost optimization has always been a pain for cloud users in general. In 2019 AWS introduced a few welcome services that try to address the problem.

Amazon EC2 Resource Optimization Recommendations offer predictions based on historical compute usage (among other things) that help users choose the right instances for their usage. Along with AWS Compute Optimizer, which also tells you when you should upgrade to a higher instance type, these two are a step in the right direction.
Budget Reports make reporting easier and help set budget alarms for different scenarios.
AWS Solutions are an attempt to provide best practices and solutions to common problems for everyone. These solutions are created and/or vetted by AWS Architects and are, by definition, perfectly optimized for their particular scenarios. Some examples include real-time monitoring of IoT devices, centralized logging, .NET Serverless CI/CD, etc.

There are a lot more features/services in this category (like AWS Backup) and it would take too much time to list them all. It’s worth noting that nearly all of them try to simplify or entirely automate tasks that users have trouble with.

There’s definitely a reaction to smaller cloud providers that offer a less confusing/cheaper experience for newcomers. AWS is taking note and acting accordingly.

Security & Governance

Cloud security – the favorite topic of technology journalists and new cloud users alike. Both don’t get how it works and adding new features/services won’t change that. Still, it doesn’t hurt to simplify things once in a while.

IAM Access Analyzer is a feature that removes tedious manual checks and helps security admins easily analyze granted permissions. It also provides useful data like “service last accessed” timestamps.
Amazon Detective is a service that’s still in preview but has the potential to be a game-changer. Right now, determining what caused a security issue means collecting data from 5-6 different services and analyzing it yourself. At an enterprise level, this gets extremely complicated. Amazon Detective will try to put an end to that by analyzing data from multiple sources and unifying it into a single view. That way, administrators can see all the data in one place, along with historical patterns without deploying new software.

Migrating Away From Oracle

AWS probably made more noise in the media about moving away from Oracle than anything else this past year. I mean, just look at the video below.

They claim to have saved tons of time, money and headaches by finally migrating. All of the hype was an obvious jab at Oracle Founder/CTO Larry Ellison, who, in an interview last year, said that Oracle will win the cloud wars among other ridiculous things.

This was an absurd thing to say as no startup wants to build their business on top of Oracle cloud. But it did get the media’s attention.

Fast-forward to a year later, Amazon migrated away from Oracle and trolled Larry Ellison massively in the process.

That’s what you call irony.

Entry in Quantum Computing

If I have to choose the next big hype train after the ML/AI settles down, it would be Quantum Computing.

And I’ll be honest – I have no idea how it works.

Still, if the biggest players in the game are working on it, it’s probably worth keeping an eye on.

Earlier in 2019, Google announced it had made a significant breakthrough.

AWS quickly followed and announced 3 main projects:

Amazon Braket – a managed services for anyone who wants to experiment with quantum computing. Braket provides a development environment to explore and design quantum algorithms. Quantum computers aren’t mass-produced yet, so these types of services are the only game in town for hobbyists.
Amazon Quantum Solutions Lab is a way for companies to get in touch with quantum computing experts. Not that exciting but really useful since people with that skillset aren’t just sitting around on Upwork.
Finally, AWS announced it was opening a new Center for Quantum Computing with the California Institute of Technology.

Quantum Computing is still in its very early R&D stage, but Big Tech is investing heavily. Don’t be surprised if it turns out to be the next big thing 5 years from now.

AWS re:Invent

AWS’s annual summit is now firmly the most significant tech event in the world. Re:Invent is a must follow for everyone, not just cloud geeks.

65 000 people showed up to re:Invent this year. The event itself was massive, with new services and features being announced every minute. A lot of them had to do with new trends in tech like 5G for ultra-low latency, machine learning for just about everything and quantum computing for whoever knows what it can do.

If you want a quick recap of the event, check out this article from AWS or AWS’s YouTube channel, if you have a few days available.

What to Expect From AWS in 2020

Last year saw AWS hold on to its lead in the market and establish itself as the platform for the future.

From the looks of it, AWS is aiming to improve in a few key spots in 2020:

Artificial Intelligence – The stage has been set for AWS to take the lead in this area. The services and specializations I already mentioned plus the hardware (AWS Inferentia) make AWS a natural choice for developers and data scientists. That said, other big players also have their sights on this market. There isn’t a clear winner in this space yet.

Quantum Computing – I don’t want to beat this point to death. AWS is investing heavily in this area, as are the other cloud providers. Next year won’t be definitive since the technology is still in its infancy, but it will show us who wants to compete in this area.

Pricing and Simplicity – big cloud vendors (AWS included) have been riding a non-stop wave of innovation and profits for more than 10 years now. As of 2020, they’re not the only game in town, though. Smaller providers snuck up and took a piece of the market. More specifically, they focused on customers that wanted to start quickly and work cheaply. AWS isn’t giving up that segment without a fight.

The new Savings Plans, Resource Optimization Recommendations, Compute Optimizer, AWS Solutions and Lightsail all try to appeal to an audience that wants something cheaper and easier to use. You can bet more features and services of this type will follow next year.

Cloud Wars Get Even More Intense – the move to the cloud isn’t slowing down. According to Gartner, 75% of all databases will be in the cloud by 2022. Forbes also predicts that 83% of all enterprise workloads will be in the cloud by 2020. You can bet we’ll see even more situations like the JEDI contract controversy.

If you don’t know about it, here’s a short summary: Microsoft won a 10 billion dollar cloud services contract with the US Government. AWS didn’t like that and is now fighting the results. Their CEO Andy Jassy publicly disputed the outcome and slandered Microsoft during re:Invent.

Expect more of the same (but worse) this year.