I am the incumbent candidate for Project Technical Lead to Nova
I was one of the original authors of nova when it was still part of the NASA Nebula project. OpenStack has been my life for the last 2 years. I was elected to the PPB during the first election, and then elected as the PTL for Nova. I have been serving as the PTL for the last year. There is a huge amount of work involved in being a PTL for such a large project, including managing blueprints and scheduling, bug triaging, mailing list debates, working with partner companies, and answering mailing list questions. In addition to these managerial tasks, I work on all aspects of the nova code base, and spend lots of time answering questions on IRC and launchpad.
In addition to my open source duties, which take up the majority of my time, I am also on the Rackspace Cloud Builders dev team. We deal with customers who are using OpenStack in production today, so I have intimate knowledge of the pain-points in getting OpenStack running in production.
I'm not running for a PPB seat specifically, but so much of what I do involves OpenStack as a whole, I decided to answer both sets of questions below.
1a. Since the last elections, what areas have you focused on and what contributions have you made in order to improve OpenStack as a whole?
I've been spending time making sure that nova is integrated well with the other projects, especially focusing on keystone.
I've also been working closely with Dan on Quantum integration and a plan for the openstack networking api.
In addition I've been working with various groups on elevating nova-volumes into its own project.
1b. Since the last elections, what areas have you focused on and what contributions have you made in order to improve your project?
Primarily I have been spending time trying to get everything stabilized and cleaned up so that nova is production-ready in essex.
I have directly implemented a number of blueprints and bugfixes.
I have also participated in a huge number of code reviews.
2a. What are the most pressing/important issues facing OpenStack as a whole?
OpenStack is old enough now that large entities are starting to deploy it in earnest. Now it is about ramping up adoption so it is ubiquitous. People who are wanting to use OpenStack are primarily concerned with running a cloud with minimal effort. To support these users we need to focus on:
- Consistency between projects
- Federation and bursting
2b. What are the most pressing/important issues facing your project?
There are two classes of issues facing nova. The first class includes issues that face operators and deployers. These are important because they drive code requirements. As a developer it is very easy to get tunnel-vision and try to solve only developer issues. I think one of the roles of the PTL is to keep the user-facing issues in mind so that they don't get lost.
This is primarily a documentation effort, but also includes dev work in improving admin-apis and tooling.
A lot of progress was made during Essex in this area. Larger scale testing and QA will help here. Some re-architecting around orchestration and state-management is probably necessary to achieve stability goals.
This is a larger issue that we will be facing over the next year. Nova needs some re-architecting to improve security. We need to address this a in phases, but it is probably the most important "Feature"
Federation is one of the selling points of openstack, but this area needs specific focus if we are ever going to move virtual machines between clouds easily
The second class of issues are code and developer related issues.
- Nova is too large
Various efforts are underway to decrease the surface area of Nova. The subteam strategy we implemented during essex was only successful in a few cases. A plan has been formed around splitting nova-volume into its own project. The replacement of nova-network with quantum/melange that is planned over the next year should similarly isolate nova-network.
The suggestion of specific public branches managed by subteams and rarer trunk merges is excellent and makes a lot of sense for certain areas, especially xen vs. libvirt for example. The details of this will require some hashing-out at the summit, but I'm all for giving smaller groups of developers ownership over specific areas of the code.
- Interfaces between components are not clear
The suggestion above should help with this. It has been mentioned before, but we have very little chance of upgrading without downtime unless we solve this issue.
3. What is your relationship to OpenStack & why is its success important to you and/or your company?
I described my history with OpenStack and day-to-day activity above, but OpenStack is also very dear to me personally. OpenStack is the most fun project I have ever worked on. It covers a very large area so there are always opportunities to learn new things. We have a wonderful community and it is always amazing to me to see contributions coming in from new people almost every day.
OpenStack is also very important to my company (rackspace). Rackspace has devoted a huge amount of time and energy to making this project a success.