Blog

Recap of re:Invent: It’s time to ride the dragon

You must be living under a rock if you missed AWS’s big announcement week: its new service offering, AWS Private 5G.

Before I get into the details, I have to say this is why we’re using AWS as the primary public cloud for Totogi, the product-driven telco enterprise software company where I serve as acting CEO. AWS just keeps crushing the other hyperscalers. I mean, we love Google Cloud Platform. We love Microsoft Azure. But AWS just keeps coming out with more stuff, better prices, and faster tech. It’s hard for anyone else to keep up.

Attending AWS Re:Invent 2023? Meet you there! Book a meeting with Totogi →
Recap of re:Invent: It’s time to ride the dragon

AWS Private 5G

With this new managed service, AWS aims to help enterprises install their own private 5G mobile network “in days instead of months.” Want one? It’s as easy as going to your AWS console, typing in where you want the network, and selecting the capacity you need. Boom. AWS delivers everything needed, I assume by backing up one of those Amazon delivery vans to your loading dock while making its special quacking-duck noise. Setup and deployment is reportedly “automated.” Upfront cost is zero, then it’s pay as you go. And get this: AWS manages the whole thing for you. It’s a service. WOW.

Just for comparison, let’s review the telco option for the same thing. If an enterprise needs a private network for its manufacturing facility in the boondocks, they call you up. You say it’s going to take at least 100 days to get up and running, leaving them to buy all the equipment and train their own team of people to run it. It’s tons of money, complicated, and a shit-ton of work. Basically: a big ol’ headache.

If you were an enterprise CTO, which would you choose?

So. What are you gonna do about it? Stick your head in the sand and complain about how you can’t trust the hyperscalers? Or are you going to finally realize you need to work with them so they can integrate with your network and offer customers easy-to-use packaged services? With enterprise being 50% of your business case for the 5G network, my bet is it’s going to have to be the latter.

Like I said in my MWC21 keynote: you gotta ride the dragon. Telcos need to expose a hyperscaler-friendly tech stack at the edge so enterprises can easily use it, with the tools and environment they are used to. More and more, that environment is a hyperscaler environment, and with 50% market share, that environment is AWS. Stop dragging your feet and climb into that saddle!

Graviton3

The AWS Graviton processors have to be some of the best innovations to come out of AWS. Amazon developed them specifically for cloud computing and in just three years has already improved them *twice.* Graviton2, released in 2019, already beat Intel’s x86-based instances by 40% on price and 7x on performance. (AnandTech’s benchmark test was so dramatic, they called it the “x86 Massacre.” Brutal, but accurate.)

Now, AWS says that compared to Graviton2, the newly announced Graviton3 will be up to 25% faster, with 2x faster floating-point performances, and a 3x speedup for machine-learning workloads. Plus, it uses 60% less power, providing way better price performance than anything else out there and a lower carbon footprint.

And AWS may be demurring with these numbers, because Liz Fong-Jones of honeycomb.io cites even more impressive performance increases.

The best part? Any AWS customer can use these chips and get this price improvement with hardly any work! You don’t have to write a business case, fight with procurement to buy new servers, or even install anything. Amazon already made the R&D investment, and you get the benefits without lifting a finger—other than moving your workloads to the new machine types.

Know that over in Totogi-land we’re moving our products to take advantage of this goodness. It’s one of the key reasons we’ve built on top of the public cloud. It innovates underneath us and we just need to be nimble enough to keep up with the new features. It’s awesome.

DynamoDB at 60% pricing

The last announcement that makes my top three list: the new Standard-Infrequent Access table class for DynamoDB. It reduces storage costs by 60% compared to Standard tables—and gives you the same performance, durability, and scaling.

At Totogi, we’re all over this. Our customers have infrequently accessed data that we need to store—so our cost just went down and we don’t have to give up a single thing (or even do anything to get it). We’ll still get single-digit millisecond read and write performance from DynamoDB with the same APIs. It’s just cheaper now.

Before last week, options for this kind of data were to delete it (not really an option), archive it in cheaper storage (which isn’t highly available), or keep paying premium rates. With the new DynamoDB Standard-IA, we’re building with this tech, so we just get the benefit without doing any engineering work (just like with Graviton3). Everyone in the world that uses DynamoDB just got this win.

AWS, you did it again. Amazon’s own massive platform and fulfillment operation constantly drive innovation. AWS simply invests and ships more—and it shows. None of the other hyperscalers have as good a chip set. None of them are dropping prices as fast.

In today’s cloud wars, the customer continues to be the winner. Sign me up!

Recent Posts

  1. Telco’s shift to the public cloud gets real
  2. Why AI’s future might depend on nuclear energy
  3. The public cloud is PERFECT for Africa
  4. Cloud burst for disaster recovery
  5. How to transform your customer support with AI

Telco in 20 Podcast - Tune in

Get my FREE insider newsletter, delivered every two weeks, with curated content to help telco execs across the globe move to the public cloud.

"*" indicates required fields

This field is for validation purposes and should be left unchanged.

More from TelcoDR