Security Cloud Privacy Tech
Monday Night Live With Peter DeSantis

Monday Night Live With Peter DeSantis

At re:Invent or attending remotely? Check out my “Ultimate Guide to AWS re:Invent 2022” post for some tips and tricks to get the most out of the show.

Monday Night Live with Peter DeSantis is always a fantastic talk. Peter talks us behinds the scenes of AWS and shows how various technical improvements are enabling the features and functionality that we know and love.

I live tweeted the keynote. Frustratingly in a split thread, but here are both parts…

First Twitter Thread 🧡

Tweet 1/3 πŸ‘‡ Next tweet

here we go! Monday Night Live with Peter DeSantis is about to kick off at @awscloud #reinvent 2022!

a πŸ§΅πŸ‘‡ (/cc @AWSEvents)

Tweet 2/3 πŸ‘‡ Next tweet πŸ‘† Start

@awscloud @AWSEvents I love how much fun Peter has with this keynote!


Tweet 3/3 πŸ‘‡ Next tweet πŸ‘† Start

Peter reminds everyone that this keynote is all about "how" @awscloud does things. lots of behind the scenes info in this one..


Second Twitter Thread 🧡

Tweet 1/71 πŸ‘‡ Next tweet

big focus on performance tonight.

Peter reminds us it’s not all about what you can see

“Great performance is the result of innovation from the ground up and investing over time”


Tweet 2/71 πŸ‘‡ Next tweet πŸ‘† Start

starting with a dive into @awscloud Nitro...

more on Nitro at


Tweet 3/71 πŸ‘‡ Next tweet πŸ‘† Start

here’s a great talk from Abby Fuller on the technical details of AWS Nitro:


Tweet 4/71 πŸ‘‡ Next tweet πŸ‘† Start

Peter announces a new version of Nitro (v5), with big gains behind the scenes. continued innovation behind the scenes from @awscloud here


Tweet 5/71 πŸ‘‡ Next tweet πŸ‘† Start



uses the new Nitro v5 chip under the hood. lots of benefits. tip of the iceberg here


Tweet 6/71 πŸ‘‡ Next tweet πŸ‘† Start



custom tuned Graviton3 silicon. lots of benefits for the targeted use cases


Tweet 7/71 πŸ‘‡ Next tweet πŸ‘† Start

NEW INSTANCE TYPE (coming soon):


…combines the Graviton3E + Nitro v5 for some big boosts for HPC applications


Tweet 8/71 πŸ‘‡ Next tweet πŸ‘† Start

not sure I’m digging the graphic style for these slides...


Tweet 9/71 πŸ‘‡ Next tweet πŸ‘† Start

typography with Amazon Ember is on point though. nice and clean. super clear. great typeface


Tweet 10/71 πŸ‘‡ Next tweet πŸ‘† Start

back to the tech, here’s a great summary of the new instance types just announce from @jeffbarr:


Tweet 11/71 πŸ‘‡ Next tweet πŸ‘† Start

btw, more on the Graviton processor at


Tweet 12/71 πŸ‘‡ Next tweet πŸ‘† Start

Peter has moved on to how the @awscloud Global Infrastructure is designed and built

specifically talking about SRD (scalable reliable datagram)

more on that at:


Tweet 13/71 πŸ‘‡ Next tweet πŸ‘† Start

more on the @awscloud global infrastructure at


Tweet 14/71 πŸ‘‡ Next tweet πŸ‘† Start

Peter is giving a great high level overview of TCP as a protocol

tl:dr it works great, but focuses on β€œgetting there” vs. performance

at @awscloud scale, they need something more efficient and focused on hyper scale issues


Tweet 15/71 πŸ‘‡ Next tweet πŸ‘† Start

key benefits of SRD on the @awscloud network:
  • multi-pathing
  • retries in microseconds
  • runs on nitro (dedicated hardware)


Tweet 16/71 πŸ‘‡ Next tweet πŸ‘† Start

remember, @awscloud runs at a completely different scale. very few companies run at this size. it’s a completely different set of problems, requiring completely different solutions


Tweet 17/71 πŸ‘‡ Next tweet πŸ‘† Start

SRD is helping deliver better performance for Amazon EBS, especially for β€œtail latency” …average latency doesn’t matter if you just need the data!


Tweet 18/71 πŸ‘‡ Next tweet πŸ‘† Start

teal line is SRD performance. big improvement overall, but massive in the edge cases

…btw, all new EBS io2 volumes will get this performance boost in the new year (stay tuned, they are just saying β€œcoming soon")


Tweet 19/71 πŸ‘‡ Next tweet πŸ‘† Start


ENA Express, brings SRD to ENA…available today!

more details on ENA (basic, not express):


Tweet 20/71 πŸ‘‡ Next tweet πŸ‘† Start

here’s the SRD boost for ENA with ENA Express (needs to be enabled on your ENAs)


Tweet 21/71 πŸ‘‡ Next tweet πŸ‘† Start

lots of improvements backed by the move to SRD in the @awscloud network


Tweet 22/71 πŸ‘‡ Next tweet πŸ‘† Start

β€œA veritable hoedown of machine learning”, Peter DeSantis


Tweet 23/71 πŸ‘‡ Next tweet πŸ‘† Start

oh Peter….


Tweet 24/71 πŸ‘‡ Next tweet πŸ‘† Start

this graphic redeems part of the hoedown comment…

would have been better if he drove it out on stage πŸ˜‰


Tweet 25/71 πŸ‘‡ Next tweet πŸ‘† Start

gist of this section: 🧠 machine learning models are getting MASSIVE

they need more performance (processing, storage, network, etc.) in order to train properly. @awscloud innovations in performance engineering like SRD are helping drive these advances


Tweet 26/71 πŸ‘‡ Next tweet πŸ‘† Start

here are the details on the @awscloud Trn1 instance type that Peter referenced:


Tweet 27/71 πŸ‘‡ Next tweet πŸ‘† Start

while they make the point, graphs without units and numbers are tough for me to look at

that said, numbers and more specifics would actual make these slides weaker and muddle the point Peter is making


Tweet 28/71 πŸ‘‡ Next tweet πŸ‘† Start

β€œJust as a refresher. Stochastic rounding is probabilistic rounding”, Peter


Tweet 29/71 πŸ‘‡ Next tweet πŸ‘† Start

why is that important πŸ‘†? it helps train the more faster with less loss. @awscloud accomplishes this with hardware acceleration in the Trainium silicon

more on that chip at


Tweet 30/71 πŸ‘‡ Next tweet πŸ‘† Start

also, here’s a mathematical primer on stochastic rounding:

you know, because MATH! 🀣


Tweet 31/71 πŸ‘‡ Next tweet πŸ‘† Start

the tl:dr of this section is that @awscloud has been focusing on a lot of deeply interesting (to me at least!) performance engineering problems so that we can build & train ml 🧠 models faster and more accurately


Tweet 32/71 πŸ‘‡ Next tweet πŸ‘† Start

how else would we get fun/weird avatar pics in different styles?



Tweet 33/71 πŸ‘‡ Next tweet πŸ‘† Start

Peter is currently giving a great explanation of how to scale out a workload across multiple processors. I can’t explain it in 280 character chunks 🀦

basically: many hands make light work?


Tweet 34/71 πŸ‘‡ Next tweet πŸ‘† Start

you’re going to have to watch this segment for yourself. it starts at 35m into this keynote with this slide πŸ‘‡


Tweet 35/71 πŸ‘‡ Next tweet πŸ‘† Start

Peter brings this back home using a Rings of Power reference. btw, the show is very much worth watching


Tweet 36/71 πŸ‘‡ Next tweet πŸ‘† Start

lots of work to average out the results, but the β€œring of rings” works much faster than β€œone ring”

…as long as you don’t give 9 to the race of man, 3 to the Elves, 7 for the Dwarves, and one to Sauron 🀦


Tweet 37/71 πŸ‘‡ Next tweet πŸ‘† Start

new instance type coming soon to help speed things up even more:


  • faster networking
  • faster training


Tweet 38/71 πŸ‘‡ Next tweet πŸ‘† Start

Jock Clear up now talking about the technical complexity of F1 racing

more details on @awscloud + F1 at


Tweet 39/71 πŸ‘‡ Next tweet πŸ‘† Start

three key aspects of making zoom-y cars zoom faster...

(somewhere @mosescj58 just felt a chill 🀣)


Tweet 40/71 πŸ‘‡ Next tweet πŸ‘† Start

with @awscloud, Jock and the team built a virtual ground speed sensor

it’s an AI/ML tool that provides more accurate data faster…and more importantly with less weight

…would love to see more technical details on that one


Tweet 41/71 πŸ‘‡ Next tweet πŸ‘† Start

btw, @QuinnyPig is also live tweeting this keynote. honestly, probably with a lot moreβ€”and funnierβ€”jokes

check out his thread at


Tweet 42/71 πŸ‘‡ Next tweet πŸ‘† Start

Jock continues to lay out the challenges that F1 teams face outside of the race and the onslaught of challenges within a race

lots of engineering challenges and trade offs…


Tweet 43/71 πŸ‘‡ Next tweet πŸ‘† Start

I’m not a car person (see the previous β€œzoom-y” comment), but I do love engineering challenges. lots of crazy ones here

Jock is constantly calling out the sheer number of decisions the team needs to make. they need data & analysis to make those decisions


Tweet 44/71 πŸ‘‡ Next tweet πŸ‘† Start

they are launching an app with a bunch of race/interview content

…was hoping for an engineering data app like the real time MLB analysis:’s not that, but I’m sure F1 fans will enjoy it


Tweet 45/71 πŸ‘‡ Next tweet πŸ‘† Start

Peter is back up now. new topic: #serverless computing


Tweet 46/71 πŸ‘‡ Next tweet πŸ‘† Start

β€œBehind me you can see a youthful @Werner announcing @awscloud Lambda”, Peter



Tweet 47/71 πŸ‘‡ Next tweet πŸ‘† Start

Lambda has grown a LOT since it’s launch in 2014

thanks again @timallenwagner for setting us on this path!


Tweet 48/71 πŸ‘‡ Next tweet πŸ‘† Start

Peter is diving into the technical details and some of the performance engineering that goes into @awscloud Lambda in order to reach this scale...


Tweet 49/71 πŸ‘‡ Next tweet πŸ‘† Start

if you are interested in not only #serverless compute, but the whole concept. check out


Tweet 50/71 πŸ‘‡ Next tweet πŸ‘† Start


Tweet 51/71 πŸ‘‡ Next tweet πŸ‘† Start

Peter using a castle 🏰 analogy for virtual machines. key point: small number of access points. built for defence


Tweet 52/71 πŸ‘‡ Next tweet πŸ‘† Start

β€œMostly secure isn’t good enough”, Peter


Tweet 53/71 πŸ‘‡ Next tweet πŸ‘† Start


Tweet 54/71 πŸ‘‡ Next tweet πŸ‘† Start

β€œAt @awscloud, we always use an isolated virtual machine for customer compute workloads”, Peter >> thus Nitro!


Tweet 55/71 πŸ‘‡ Next tweet πŸ‘† Start

back to Lambda…

think of it as a cache. a big pool of compute


Tweet 56/71 πŸ‘‡ Next tweet πŸ‘† Start

each slot was specific to a customer function. new function? new virtual machine

if a VM is up for the customer function, great. if not, initialize a new instance (empty & waiting) to handle it >> a/k/a cold starts

…it works, but there’s room for improvement


Tweet 57/71 πŸ‘‡ Next tweet πŸ‘† Start

more graphs


Tweet 58/71 πŸ‘‡ Next tweet πŸ‘† Start

so cold starts are a cache miss. that happens, caching isn’t perfect


Tweet 59/71 πŸ‘‡ Next tweet πŸ‘† Start

obvious solution: bigger cache on the @awscloud backend

that’s more expensive and less efficient

enter Firecracker


Tweet 60/71 πŸ‘‡ Next tweet πŸ‘† Start

Tweet 61/71 πŸ‘‡ Next tweet πŸ‘† Start

moving to Firecrack provides more efficient use of the existing hardware. it also spins up faster, allowing things to sit in the cache longer!


Tweet 62/71 πŸ‘‡ Next tweet πŸ‘† Start

results are Firecracker, bigger arrow, less bump

numbers would be helpful here. Peter mentioned one but I’m not sure I heard it right

assumption: things got better with the move to Firecracker


Tweet 63/71 πŸ‘‡ Next tweet πŸ‘† Start

even with these improvements, you’ll see still cache misses (a/k/a cold starts)

need to make them faster. where is that time going?

for Java devs? a lot of it is in initialization


Tweet 64/71 πŸ‘‡ Next tweet πŸ‘† Start

solution is to save the initialized start. think opening up your MacBook’s lid and working vs. turning it on from completely off


Tweet 65/71 πŸ‘‡ Next tweet πŸ‘† Start

this feature is called AWS Lambda SnapStart and it’s available today!

bye bye cold starts…at least most of them


Tweet 66/71 πŸ‘‡ Next tweet πŸ‘† Start

places where SnapStart _won’t_ work/help #1:
  • some cryptography use cases where uniqueness is required
  • can adjust with code changes

#reinvent #serverless

Tweet 67/71 πŸ‘‡ Next tweet πŸ‘† Start

interesting division of snapshot-ing here. separating out the uniqueness to improve performance even further


Tweet 68/71 πŸ‘‡ Next tweet πŸ‘† Start

last piece of the puzzle, predictive snapshot loading

get the pieces needed to start the function up and running as quickly as possible…before it’s even asked for


Tweet 69/71 πŸ‘‡ Next tweet πŸ‘† Start

pumped to see the benefits of @awscloud Lambda SnapStat. love that I really don’t have to do anything to get those benefits

#reinvent #serverless

Tweet 70/71 πŸ‘‡ Next tweet πŸ‘† Start

Peter signs off with some great session recommendations for performance engineering


Tweet 71/71 πŸ‘‡ Next tweet πŸ‘† Start

great keynote from Peter DeSantis.hoping we’re going to see some @AmazonScience or @awscloud papers on these innovations soon

lots of great advancements & engineering work to deliver more features & functionality to builders

…more tomorrow as #reinvent kicks into high gear!

More Content