Andrei Kaigorodov

I used to attend their watch parties in Berlin with a live stream from Las Vegas, but this year AWS doesn't seem to be hosting one in Berlin

Saying 197, it is well, not entirely new products but one feature here, a new region there, a billing change, and so on. Still, I guess there will be some big announcements during the conf. An engineer myself, I can only feel compassion for their engineering teams now trying to meet the deadlines

As re:Invent 2024 approaches, the number of updates from AWS is overwhelming. Not surprising, given their sheer size, but still. This post go.aws/49913G1 says that they've had 197 new launches

#AWS #AWSreInvent

and a somewhat informal conclusion for a research paper: "do not use Python" 😀 though it doesn't seem to be published anywhere except of Research Gate

so, neither C Libwebsockets nor Python's websockets libs could even complete the benchmark:

The diagram showing the use of additional TM pods acting as stand by

interesting detail is the use of standby task managers. For high availability, usually you set up only stand-by job manager, as task managers can be recovered (processing might take a hit though). They also pre-provision task managers to make recovery faster if one of task managers pod fails #flink

Building a User Signals Platform at Airbnb How Airbnb built a stream processing platform to power user personalization.

An interesting read about stream processing platform for user personalization built by Airbnb bit.ly/3CEQVsF that is based on #Kafka and #Flink

Netflix’s Distributed Counter Abstraction By: Rajiv Shringi, Oleksii Tkachuk, Kartik Sathyanarayanan

There are plenty of people writing about the recent Netflix outage, instead I’d like to share their latest engineering blog post bit.ly/3ZbIyxp about distributed counters. Interestingly, they didn't just explain their design,but went through a few approaches and their pros & cons #distributedsystems

"A distributed system is one in which the failure of a computer you didn't even know existed can render your own computer unusable," said Leslie Lamport - one of the pioneers who laid the very foundations of distributed system theory
#distributedsystems #LeslieLamport

Fallacies of Distributed Computing

1 The network is reliable
2 Latency is zero
3 Bandwidth is infinite
4 The network is secure
5 Topology doesnt change
6 There is one administrator
7 Transport cost is zero
8 The network is homogeneous

formulated in 90s but still relevant #DistributedSystems