Udo Seidel is no stranger to dealing with enormous file stores. Udo is Section Manager for Linux Strategy and Server Automation at Amadeus Data Processing GmbH and will be giving a talk with Vijay Bellur at DevNation in San Francisco, California later this month on developing applications with GlusterFS and OpenStack Swift.
We sent Udo a few questions in advance of his talk to give us a little bit of a preview of what we might hear. For a little background, the DevNation agenda gives summary: "GlusterFS is a general purpose, scale-out file system that can be accessed through file, block, and object protocols. OpenStack provides services that enable block, object, and file interfaces. Together, GlusterFS and OpenStack are well-integrated for addressing several real-world use cases." For those interested in hearing Udo tell a little more, consider listening to this podcast conversation between Udo and Richard Morrell, Cloud Evangelist and Cloud Security Architect at Red Hat.
In this interview, I asked Udo about his involvement in the open source community, cloud storage, and his talk at DevNation.
Tell us how you got involved with open source.
Actually, I got there because I am lazy. Back in the nineties I was writing C program as part of my master thesis. I was using MS Windows as a platform back then. At a certain strage I hit the memory segmentation challenge (Gate A20). I had two options: either I optimize the code in regards memory allocation and de-allocation, or I switch to a platform which does not suffer from the limitation. I switched to Unix at the university and was told that Linux is "the Unix for home." I was amazed by the openess and options Linux provided to me.
What will you discuss at the DevNation conference?
I will talk about how our data centre is challenged by the paradigm change that comes with Openstack and GlusterFS. Both things are changing the game of a traditional date centre.
Tell us more about the open source projects you with in relation to Swift/GlusterFS.
I used to be interested in shared storage, mainly shared disk-based clustered file systems. I am very attracted to distributed file systems, and I should be more precise and say: distributed storage solutions. This interest was not related to Swift or OpenStack at first.
What role do storage technologies play in cloud infrastructure; in particular, how do they interact with technologies like OpenStack?
Storage is a key component of cloud infrastructure. Regardless of the stack used at the end, data has to be stored somewhere. Rephrased, one could say: Storge is part of the instructure of XaaS. Storage can (and has to) provide different services for the cloud: storage of operating system images, block storage for the compute nodes, and object store. OpenStack users will probably have identified Glance, Cinder, and Swift here. From cloud users perspective the storage technology as such should not matter as long as the requirements are met. Replacing product A with B should be transparent if not even invisible to the end user.
What aspects of open source do you apply to your interests outside of technology?
I enjoy the sharing of knowledge, inviting others to join a task/mission, acting as a team, and openess (in the sense of honesty and directness).
1 Comment