Scaling Systems at @AWS | PhD survivor @LSDSImperial @imperialcollege | Opinions and retweets my own
San Fransisco, Bay Area
Joined October 2011
- Tweets178
- Following414
- Followers268
- Likes1.1K
Pinned Tweet
Great summary of #MEDEA our recent work with Microsoft on Apache YARN by @adriancolyer at #themorningpaper ! teapawt.net/adriancolyer/status/10…
"Medea: scheduling of long-running applications applications in shared production clusters" Garefalakis et al., EuroSys'18. blog.acolyer.org/2018/06/13/… #themorningpaper
If you run clusters with a mix of batch and long-running (e.g., web) containers, then check out Medea...
Panagiotis Garefalakis retweeted
Please retweet!
Searching for a PhD student for my VIDI grant. The team is building a programming model & self-driving streaming dataflow engine for Transactional Stateful Functions.
Apply here by 10th February (and drop me an email): tudelft.nl/over-tu-delft/wer…
Panagiotis Garefalakis retweeted
A year ago, @immerokcom 🐿️s joined @confluentinc, and the team has already shipped an open preview of @ApacheFlink serverless offering. 😍
Join our stream processing workshop in less than a week to see what we've built so far. 🤓🏗️
events.confluent.io/ws-strea…
Let me know if you are attending #FlinkForward and want to chat more about @ApacheFlink and how we are dealing with streaming failures at @confluentinc !
Nov 8th is right around the corner! Join @confluentinc to learn how to prepare for and handle failures in your dataflows using #ApacheFlink in this tech deep dive session.
🔩Become a technical pro at #FlinkForward with training prior to the conference: bit.ly/3QiEtD3
Panagiotis Garefalakis retweeted
On Tues Nov 7, we have TWO meetups happening. First up, a social mixer hosted by @Onehousehq and @confluentinc for folks attending #flinkForward.
meetup.com/seattle-data-engi…
Panagiotis Garefalakis retweeted
🐿 The open preview of @ApacheFlink on Confluent Cloud is now available!
Now, you can easily build high-quality, reusable data streams with the industry's only cloud-native, serverless Flink service.
Check out our blog to see how it works ➡️ cnfl.io/463JEfF | #Current23
Panagiotis Garefalakis retweeted
Here we go, ready to disrupt the industry. The most exciting product I’ve ever worked on is entering the Open Preview! 😍 #Current23
Panagiotis Garefalakis retweeted
FLIP-291 brings the Adaptive Scheduler one step closer to becoming the new default for streaming and marks an essential milestone in building the state-of-the-art AutoScaler for @ApacheFlink 🐿️🤓
Read more in a 🧵👇
Panagiotis Garefalakis retweeted
This week’s podcast is framed as a Flink vs Kafka Streams vs ksqlDB debate, but if you ask me the winner is anyone who learns some new stream-processing techniques. 😀
Comparing Stream Processing Tools youtu.be/Wqko7MunKZs with @MatthiasJSax and @jwfbean.
Panagiotis Garefalakis retweeted
Fancy a physical (!) trip to London to discuss your work and network with the British database community this December at @imperialcollege? Consider submitting a paper to @BICOD2021: bicod.uk/cfp. Despite the name, we invite contributions from anywhere.
Panagiotis Garefalakis retweeted
nice to see Apache Hive getting a product placement here
This quoted post is unavailable.
Panagiotis Garefalakis retweeted
On January 20, the @ApacheCalcite community is organizing a virtual meetup (meetup.com/Apache-Calcite/ev…); if you are interested in data management & data integration check the agenda and join the event :)
Panagiotis Garefalakis retweeted
You are in a job interview. The interviewer says "Do you understand Kerberos?". What do you say?
21%No
7%Yes
54%Define 'undertand'?
18%This interview is over
101 votes • Final results Panagiotis Garefalakis retweeted
Please join us Thursday (5/11) at 2pm UTC for Marios Kogias', Microsoft Research Cambridge, online talk "Making RPCs first class datacenter citizens" lsds.doc.ic.ac.uk/content/ma…
Thanks for attending! Slides available below:
slideshare.net/PanagiotisGar…
An online presentation about runtime filtering enhancements in @ApacheHive tomorrow at 20:15 GMT+2 @ApacheCon along with @pgaref . Pass by to say hi and possibly learn something new 😄
apachecon.com/acah2020/track…
Panagiotis Garefalakis retweeted
An online presentation about runtime filtering enhancements in @ApacheHive tomorrow at 20:15 GMT+2 @ApacheCon along with @pgaref . Pass by to say hi and possibly learn something new 😄
apachecon.com/acah2020/track…
Excited to be giving a talk @ApacheCon this Tuesday about the Hive runtime filtering enhancements I have been working on for the past few months with some neat perf results! Join to learn more:
apachecon.com/acah2020/track…
#ACAH2020 #OpenSource #BigData #ApacheHive #ApacheOrc
Panagiotis Garefalakis retweeted
Unless you care about latency.
You’re unable to view this Post because this account owner limits who can view their Posts. Learn more
Panagiotis Garefalakis retweeted
If you liked our #SIGMOD2020 tutorial on the evolution of stream processing systems, our extended (30+ pages) survey on the topic is now online: arxiv.org/abs/2008.00842
Feedback and comments welcome!
Join us tomorrow at #SIGMOD2020 @SIGMOD2020 for a 3h #tutorial on the evolution of stream processing systems.
We start at 10:30am PDT.
Slides are already available on our website: streaming-research.github.io…
(/w @SenorCarbone @MarioFragkoulis and @kAsterios' blessing 🙏🏼)
Panagiotis Garefalakis retweeted
TIL that k8s memory limits are bad for processes which use mmap() - the memory.usage_in_bytes includes all pages, not just anonymous pages.
Been open for a few years now
github.com/kubernetes/kubern…
Deja vu, back to YARN-1775
Panagiotis Garefalakis retweeted
Our paper "LightSaber: Efficient Window Aggregation on Multi-core Processors" was accepted in SIGMOD'20! Check out the code of #LightSaber (github.com/lsds/LightSaber)! Kudos to @akoliousis @ppietzuch and @holger_p. @ICComputing @LSDSImperial
