Despite the virtual format, this year’s AWS re:Invent was as pivotal as ever, offering three weeks of announcements, content, resources, and training at no cost. This article breaks down the major highlights, announcements, and key takeaways you need to know from AWS re:Invent 2020.
I never thought I’d say this, but after three straight weeks of logging into virtual sessions for AWS re:Invent 2020, I’m looking forward to the next real-life conference or trade show I attend. Bring on the bad coffee, achy feet, and waiting in line. Yes, I said it. I miss all the in-person conference hooplah!
But don’t get me wrong. Despite the virtual format, AWS did a great job making this year’s re:Invent as pivotal as ever. There really was something for everyone, with dozens of product/service launches and over 50 different tracks for technical folks to dive deep into relevant topics, AWS leadership sessions that inspire innovative roadmap planning, and a Global Partner Summit to foster relationships with AWS Partners who can help round out your cloud needs. (Ahem. CloudHealth is an AWS Technology Partner!)
Culture of innovation
Appropriately, CEO Andy Jassy kicked off his keynote talking about AWS’ culture of innovation and how to reinvent yourself as a business. Jassy stressed the need for businesses to reinvent themselves all the time, and never wait until you’re becoming extinct to do so. He shared the popular example of Netflix cannibalizing its own DVD service with streaming, saving itself from future despair by innovating the in-home TV experience.
With AWS’ 29% year-over-year growth, businesses would be foolish not to heed lessons from the cloud business behemoth. My favorite lesson relates to speed—Jassy adopts Jeff Bezos’ practice of not wasting time deliberating over decisions that are easily reversible, or “have two-way doors.” The only decisions you should spend significant time weighing are those with “one-way doors,” or those that are not easily reversible. Don’t be afraid to fail, and if you do, fail fast, learn from your mistakes, and keep going.
AWS compute is getting (a lot) cheaper
One of the high-level takeaways from re:Invent this year is that AWS has significantly reduced the cost of compute power. Their new custom silicon chip called Graviton2 powers a new generation of EC2 instances at a 40% lower cost than previous generations. Using new instance types with this CPU not only shaves costs directly off your EC2 bill, but this custom silicon also makes it much cheaper to use compute-heavy services from AWS—like those that leverage machine learning—saving you even more money on your total cloud bill.
CloudHealth can help you understand what your EC2 usage looks like from the business context that’s important to you, whether that’s by team, business unit, account, etc., and how it’s evolved over time at an hourly, daily, weekly, or monthly level. And, of course, how the changes in usage have affected your costs.
If you’re running serverless, don’t worry—you’re not missing out on savings. Since the billing increment for AWS Lambda is changing from 100 milliseconds to one millisecond, your compute costs will go down too (saving some businesses up to 70% on Lambda costs). With this change, AWS is better delivering on their promise to only charge you for what you use. CloudHealth already supports this change to Lambda billing, so lower costs on your AWS bill will be reflected in the CloudHealth platform.
Cheaper, more flexible block storage
Amazon Elastic Block Store (EBS) got an upgrade from gp2 to gp3. What this means for you is a 20% lower cost per gigabyte than gp2 storage. This direct cut off the cost is great, but it’s just the beginning. Amazon EBS gp3 volumes allow you to provision performance (or IOPS) independent of storage capacity.
Previously, developers had to provision block storage volumes that met both the performance and storage needs of the workload, often resulting in overprovisioning for one of the two. Now, users can scale IOPS and throughput without needing to provision additional block storage capacity, and they pay only for the resources they need.
Container services “anywhere?”
Leading up to AWS re:Invent 2020, we predicted that AWS would start warming up to the term “multi-cloud.” We were sort of right.
While no one from AWS uttered the word “multi-cloud” even once during the entire three-week conference, AWS quietly announced two new services coming in 2021 that usher them into multi-cloud management: Amazon ECS Anywhere and Amazon EKS Anywhere. Jassy announced these services as ways to “run ECS or EKS in your own data center,” but didn’t elaborate much on the question we all had in our minds:
“It’s not called EKS Data Center. It’s called EKS Anywhere. Does that really mean anywhere? Could it mean… Azure? Or GCP?”
Jassy talked a lot about bare metal, virtual machines, and using this service as you transition from on-premises to the cloud. Then he offered this nugget: “You bring us compute, we don’t care where it’s from.”
You cheeky little devil.
Whether or not AWS decides to market Amazon ECS Anywhere and EKS Anywhere as “multi-cloud” remains to be seen, but the capability to run these container services on any infrastructure is there. And this strategy of supporting containers on others’ infrastructure is a massive shift for AWS. It signals their acknowledgment that other cloud providers will continue to be part of AWS customers’ ecosystems.
Living up to its ideals
If we think about AWS as a company, it’s not surprising that they’ve reinvented their core cloud offerings to be faster, cheaper, and better serve the customer. So long as AWS continues to live up to its ethos, remains customer-centric, and isn’t afraid to reinvent itself from time to time, they’ll continue to be the leading cloud provider on the globe.
And to learn more about building a successful cloud practice with AWS, see our eBook: How to Accelerate Your Cloud Journey to Reach Cloud Maturity