The OpenStack Summit, now in its 16th iteration, took place last week in Vancouver, BC. OpenStack operators and developers from around the globe descended on the beautiful Canadian city made famous for the 2010 Winter Olympics to discuss all things Cloud Open Infrastructure, and I was fortunate enough to attend for my fifth summit (Tokyo, Austin, Barcelona and Boston previously).
Memset has spent years building services on top of OpenStack the world’s biggest open infrastructure platform because we believe the future of cloud technology should be an open one. And, OpenStack’s principles of the four O’s; Open Source, Open Design, Open Development and Open Community, fall directly in line with Memset’s values, so it is a great fit.
In a closed world, very few benefit from proprietary software, which is why we have been contributors to the community since 2012 and why it powers our Cloud IaaS platform. So as Product Manager, I was once again sent off on a jet plane to attend the summit and as usual, blown away by the passion and dedication of this worldwide community, as after all, without them, I wouldn't be where I am today.
A personal highlight of mine from the Summit was a session by John Dickinson, PTL of the OpenStack Swift Project, “Swift Project Updates”, as this was mine, and Memset’s first dive into OpenStack. John has been working on the project since its inception; he talked about the past, present and future of Swift, with such passion and admiration for the community that built it, along with acknowledging its recent 8th Birthday! Happy Birthday, Swift!
OpenStack Swift is the object storage project that is designed to store blobs of content, whether that be images, text or binary data, using a simple API driven interface. John explained that the world, and its data storage requirements are vastly different to when the project started 8 years ago, and that due to this, they have recently been re-evaluating Swift’s priorities.
Storage is growing exponentially, and with that more and more storage space is required. Good timing then, for the upcoming improvements in Swift allowing for ‘zero outage’ migrations to erasure coded object storage over traditional less economical replication. This will offer ease of migration towards much higher storage efficiency, which will, in turn, allow us to save some pennies, and pass the cost savings on to our customers.
Along with this, there are many other improvements in the Swift space, John stated: “This release is as large by line count as the last ten releases combined”. In fact, interestingly 10% of the code base is new for deep bucket support, or “sharded buckets”. This allows for an arbitrary (> billions!) of objects per container, which up until now was technically possible, but was limited due to significant overheads in the SQLlite database that stores the container metadata.
Memset’s goal is to provide high performance, resilient compute services at transparent and competitive prices using Open Source software. By utilising OpenStack we are able to achieve this faster than ever before, and improvements like these by the community are key to that goal.
As Product Manager of Memsets Cloud IaaS service, the OpenStack Summit and PTG are absolutely vital to helping shape the future roadmap of our products, and further than this, gives me the opportunity to influence it. If you’re interested in joining, the next OpenStack Summit takes place in Berlin this November; maybe I’ll see you there!