Chameleon Changelog for December 2023

Happy New Year from Chameleon! A year has ended and another fabulous year begins – as always, we are looking forward to working with you and jointly developing an understanding of what new opportunities in science require what new experimental capabilities  that need to be supported by Chameleon. Before we start, let’s review how Chameleon changed, not just last month but throughout the last year. 

For the core testbed, CHI@TACC added 72 Ice Lake nodes, the CHI@UC Xilinx hardware was improved with a debugging cable. Our edge computing site CHI@Edge got 18 new Raspberry Pis to keep up with demand, and all devices received kernel upgrades. For disaggregated hardware fans, coming soon is new GigaIO hardware, similar the the Liquid hardware we already have at TACC. We also connected TACC to the FABRIC testbed and since UC had already been connected to FABRIC in the previous year, you can now run Chamelon-FABRIC experiments from both sites – this year we are planning to make those experiments a lot easier to run so watch this space for new developments! 

In addition to new hardware we also improved the CHameleon Infrastructure (CHI). You asked for many specific usability improvements; in response we streamlined the Horizon web dashboard, the reliability of reserved resources and making leases was increased, and we made it easier to manage large education projects. We also implemented additional features in Trovi, our research artifact repository. We updated the artifact home page to feature popular artifacts, added artifact badges on our dashboard to showcase both education courselets and reproducible experiments, and added metrics of artifact usage so you can tell at a glance how much impact your artifact has had. 

In other news, we added a supported Ubuntu 22.04 appliance, and updated our image build process to better support custom images. We also improved multi-node instance launches, so that launches of a large number of nodes should work much more reliably and efficiently than they used to.  

CHI-in-a-box, the packaging of CHameleon Infrastructure (CHI), improved as well to reflect improvements to the system as well as add operator tools that make resource management easier. As many of you know, CHI-in-a-Box allows others to create volunteer Chameleon sites. We would like to give a big New Year’s shoutout to operators of Chameleon sites at IIT, NCAR, Northwestern, and UIC – the additional hardware they provide supplements the Chameleon offering and adds a distributed dimension to the testbed! CHI-in-a-Box has also been used to add ephemeral resources to the testbed to support specific initiatives such as the IndySCC student competition this year – big thanks to the UC Boulder team whose resource is still online, available to all Chameleon users till Monday!

Lastly, the last year has been great for connecting with the Chameleon community. The biggest opportunity here was the Chameleon User Meeting in May with special emphasis on education, but we also organized multiple BoFs and reproducibility hackathons at USENIX FAST, and reproducibility hackathons at ACM REP, USENIX OSDI/ATC, IC2E, and one virtually in December. It was great to meet so many of you – stay tuned for the announcements, we are planning more events this year! 

Last but not least, make sure to check out our Used in Research page showcasing all of the exciting research that our users do on Chameleon – make sure to report your research so we can showcase it on our web page: we are proud to support the research you do but it has also practical value in the it lets NSF (our sponsor) know that the infrastructure is used – and useful!

We are looking forward to seeing what you can do in 2024! Happy experimenting!

Add a comment

No comments