Managing Performance in a Fast Growing, Large Scale, Private Cloud
Keynote of TestIstanbul 2023
This presentation will cover challenges of understanding performance, finding, managing and preventing performance problems and managing capacity in a fast growing private cloud. Almost all of the examples and learnings have come from my time as a performance and capacity engineer at Facebook during the 2010 – 2019 period. This was a period of rapid growth during which Facebook grew to serve more than two billion users, acquire multiple large-scale successful companies (Instagram, WhatsApp, Oculus,) expanded its private cloud from a single leased space in a shared data center to multiple data center locations around the world and started producing its own hardware (servers and networking gear).
During this time, the capacity and performance engineering team had to deal with daily challenges related to deploying new products and features, while simultaneously guaranteeing performance of existing products to its users, managing multiple different generations of servers and adjusting for multiple world-wide shortages of different hardware components or catastrophic failures of some components.
The range of topics covered in this presentation will include: performance testing on live traffic and systems, performance monitoring and analysis, debugging performance issues and performance planning for risk mitigation.
About Goranka Bjedov
Known for her technical achievements and her refreshing wit, Goranka Bjedov worked as a performance and capacity engineer at Facebook where she was responsible for making sure there are enough servers to handle everything Facebook users want to post, upload, find or otherwise engage with. Most of her time was spent analyzing performance and assessing risks. Her industry career also included a performance engineering position at Google, Network Appliance and AT&T Labs. She is a frequent keynote speaker at performance related conferences and workshops.
Prior to joining industry, Goranka was a tenured faculty at Schools of Engineering at Purdue University, teaching mostly programming classes and conducting research in large-scale computer parallelism. She co-authored two textbooks and numerous papers, and was a publication chair for Frontiers in Education conference for a decade. In that role, she co-wrote the software for the first complete Web-based conference proceedings production in 1995. She has served on the Anita Borg Scholarship committee while at Google and has been 2014 Grace Hopper Celebration Software Engineering Track Co-Chair and 2015 Grace Hopper Celebration Career Track Co-Chair.
Right now, her time is spent walking and biking different trails around the world. And, when she is back in Silicon Valley, she talks to and consults/advises a number of different startups, working in areas of machine learning, mobile development, performance engineering and health related advances.