
What we’re about
BayLISA is the premiere system administration user group in Silicon Valley and the San Francisco Bay Area. Founded in the very early 90s after the fourth Large Installation System Administration (LISA) conference, BayLISA has supported and educated systems, network, storage, virtualization, and other technology professionals in the Bay Area for over 20 years.
We use Meetup to coordinate meeting attendance, announcements, and reminders. We have our official organizational presence on the web, including links to membership, sponsorship, and mailing lists, at [www.baylisa.org](http://www.baylisa.org/).
Our legacy description:
One way to put it is that there are many user groups, but we are the sysadmins group. BayLISA includes system and network administrators across a range of skill levels. BayLISA meets monthly to discuss topics of interest to administrators and managers of sites supporting more than 100 users and/or computers. The meetings are free and open to the public.
BayLISA grew out of an after-hours discussion among attendees of the USENIX LISA IV conference. The idea was to provide a forum for Sysadmin professionals in the San Francisco Bay area to get together and exchange ideas, hear speakers address topics of interest and most importantly, socialize.
BayLISA stands for: Bay Area Large Installation System Administrators
See you at the meetup!
Upcoming events (3)
See all- Modernizing Caching at Databricks: Low Latency, Scalability & EfficiencyKoll Oakmead Park, 3350 Scott Blvd Building 54, Santa Clara, CA
Note: please fill out this simple form in advance of attending the meeting to allow for smoother flow at the front door, and help us with organizing this and upcoming events. Thank you!
Efficient caching is essential for scalable, high-performance infrastructure, yet distributed caching in data centers often lacks
well-designed architectures. At Databricks, they've been modernizing service architectures with in-memory stateful services and caching to enhance reliability, efficiency, and scalability — while reducing operational overhead.In this talk, they will introduce two key building blocks for low-latency, scalable distributed systems: Dicer and Snappy. Dicer is an auto-sharder that dynamically redistributes key assignments across pods, addressing operational challenges like pod failures, restarts, horizontal auto-scaling, load imbalances, and hot keys. Snappy ensures applications can serve mostly fresh data while eliminating the inefficiencies of TTL-based expiration and polling mechanisms.
Additionally, we’ll present Linkstore and Softstore, our linked and remote in-memory caching solutions designed to abstract caching complexities and improve system resilience. By raising the level of abstraction, these innovations empower DevOps teams to manage distributed caching more efficiently, reducing disruptions and optimizing performance at scale.
Bio:
Atul Adya is currently a Principal Software Engineer at Databricks. Previously, at Google, he was responsible for the distributed caching infrastructure systems used across the company (such as the Slicer auto-sharder and the Thialfi cache invalidation service). These systems are widely used across Google and handle billions of requests per second. Prior to Google, Atul worked at Microsoft Research and Microsoft SQL Server. He received his S.M. and PhD degrees from the Massachusetts Institute of Technology.----
BayLISA is held for system and network administrators, SRE and DevOps professionals across a range of skill levels. BayLISA meets to discuss topics of interest to this group. Our meetings are free and open to the public, and are scheduled for 7:00pm on the third Thursday of each month.
We always welcome presentation topics and volunteer speakers. Use the "Contact us" link on this page to get in touch with BayLISA's directors.