The minutes of TAB meetings are provided for informational purposes only. Discussions and decisions described in the minutes should not be interpreted as official or definitive TAB or OSDL policy. While efforts are made to ensure their accuracy and completeness, we regret we cannot guarantee that they are. If you have corrections or additions, please contact tech-board at lists.osdl.org
Friday October 6, 2006
Technical Advisory Board Conference Call Minutes
James Bottomley Randy Dunlap Tim Golden Tom Hanrahan Gerrit Huizenga Matt Mackall Leann Ogasawara Ted Ts'o Arjan van de Ven Chris Wright
1. OSDL to add Fellowship Fund fequest for TCP Testing to pending queue 2. OSDL to post Job Req for Technical Writer 3. OSDL to provide Tim Golden with information about the Fellowship Fund 4. James to get back to Tom about Adaptec Boards
1. Gerrit to address the TAB about OSDL workgroups forming a virtualization task force 2. Tim Golden - LUAC Presentation that was given to the OSDL Board 3. Review Fellowship Fund Request - TCP Testing 4. Hardware Lending Library - Adaptec boards 5. Technical Writer Status 6. TAB and community support goals for 2007 - 2009
1. Gerrit to address the TAB about OSDL workgroups forming a virtualization task force
Beginning with some background information. . . . This was originally presented to the Technical and Marketing Sub-Committee of the OSDL Board. The initial response from the Sub-Committee was that they felt this sort of thing would work against what the community wants to achieve. OSDL doesn't feel that this is the case and has asked Gerrit to come speak to the TAB to describe the 2 purposes of the task force and then have James/TAB explain if they feel there still are issues revolving around this idea or if it would be okay to proceed with bringing this topic back to the Technical and Marketing Sub-Committee.
The whole notion of creating this task force initially started as an OSDL DCL activity. There has been some confusion around the broader member companies and customer space regarding what's going on with virtualization in general. This is partly due to numerous press releases and misunderstandings. A request came in for DCL to do something. A number of proposals had been floated around and this one has had the most traction. The goal of the task force is 2 fold. Stage 1 - education for the OSDL membership on what is actually happening in the community in terms of virtualization. This will also be accompanied by the statement that the task force supports exactly what the community is doing and has no intention of going around the community. They want to work with the community. Stage 2 - A White Paper describing how virtualization gets used further up the stack. This will come as use cases from vendors and end users. How are they using it today and what capabilities exist. This will hopefully help prevent mis-set expectations. To get the widest buy in from the membership, creating a cross initiative task force with these two objectives needs to be established.
It was the impression from the Technical and Marketing Sub-Committee that Stage 1 was already satisfied with 2 article sources being sited from LWN.net and that these articles would be forwarded to the appropriate parties needing education. Questions regarding the thoroughness of the content of these articles were raised. Also, not everyone knows which sources of information are legitimate. This task force would then help not only the US members align themselves in the right direction, but also the Japanese members.
The TAB felt one of the biggest dangers is if OSDL were to give too much information that is too aggressive. It would look like OSDL is trying to steer the community, which from past experience, is something they should not do. The issue of virtualization is volatile and hostile right now and OSDL should probably not get in the middle of it. However, it was stated that there is agreement from all the members involved to help co author this paper as a statement of what they are doing. This being said, the TAB questioned why their involvement would be needed. Those members can go ahead and publish this on their own. The rebuttal was that members felt they'd want this coming from a neutral body, OSDL. They also want to avoid the paper having any sort of biased connotation and OSDL provides a neutral venue.
The TAB felt that acting as a neutral venue is something that OSDL can do but is orthogonal to the task force/initiative being brought forth today. Again raising the question of why the TAB needs to be involved. The reason given was that a status report on this very issue was originally brought to the Technical and Marketing Sub-Committee which were met with initial objections. In order for this issue to be re-addressed in front of the Technical and Marketing Sub-Committee, they need to bring back the revised proposal but also have the support of the TAB (or at least no objections).
Another question of why all the other initiatives want to be involved was raised. The answer was that mostly all the initiatives have members who are dependent on the outcome. Everyone's putting their business interest in virtualization. The TAB felt it seemed like the right thing to do there is to leave the initiative stuff out of the main document and site it elsewhere. There doesn't need to be any reference to any of the initiatives there, but is a common grounds to work together on. It's balanced and not strictly authored by just one member which the broader OSDL membership could then endorse.
The TAB felt it made sense for them to review the document the task force comes up with to ensure it doesn't misrepresent the community needs and directions. However, the TAB does not want to endorse the document. The TAB also felt that OSDL should just act as the facilitator and conduit for these member companies to come together. It was also mentioned that possibly putting a draft label on the document first and sending it to LKML before it's “quote un-quote” released might be a bridge between community egos so they can put their input in before an official version is released.
The TAB also felt that those parties involved in Stage 1 should not be involved in Stage 2, otherwise it's just plain marketing and some members of the TAB would strongly oppose this sort of action. Stage 2 needs to come from customer usage model feedback; the LUAC type of perspective. If it's just VMware and Xensource saying here's the cool ways to use our product, they can do that themselves. Target the areas that might be recommended to customers, what can be used and where. As far as the kernel and community, how you use virtualization doesn't really affect them. The TAB will remain neutral on this idea of a task force. They will neither support nor object to it. It was then restated that OSDL membership wants to go through this exercise. At some level OSDL provides them an opportunity to mutually interact and benefit. OSDL doesn't need the TAB's full support or endorsement, but they can't have their objection either.
The TAB reiterated that providing a neutral environment is something OSDL should be doing. How you use virtualization is out of the TAB's control.
Tom is going to expedite this back to the Technical and Marketing Sub-Committee.
2. Tim Golden - LUAC Presentation that was given to the OSDL Board
(Quickly had to run through the slide deck because of time.) The LUAC is struggling with the changing Linux landscape. They are wanting to transition to a more interactive advisory council rather than just a user group which gives presentations from meeting to meeting. They are the voice of the customer on where to go next. The presentation had to create a lexicon for some terms (slide 6) and an illustration of those terms can be found on slide 7. The issue of the end user and ISV's has changed. The kernel functioning is no longer what's inhibiting adoption at this point. The pain points are in up-the-stack functionality. It's the critical pieces linking back to the kernel and other OS components that enables business applications which are holding back adoption. The end users need help with the end to end computing stack. The bottom line is the LUAC wants the OSDL to address more than just the kernel. They don't expect OSDL to solve the issues themselves but rather be a guiding force and facilitator to bring the right resources together. Some of the interviews they made yielded that deployment and integration issues are slowing the pace of Linux and OSS adoption. Leading and Mid-Market are already adopting Linux, but the concern is going towards more integration. Previously, a lot of things the LUAC were focusing on technically were very binary - in other words either it works or it doesn't. Now the focus has shifted towards “this needs improvement”. Its important to OSDL because they were focusing on the binary issue but need to transition. The LUAC also perceived the OSDL goals and LUAC goals as not being aligned. Slide 11 outlines these subjective perceptions starting from kernel going all the way up to the IT community. Slide 12 was that OSDL member companies might impede effective alignment for their own purposes.
(skipping over EMEA, Japan LUAC slides) Japan is fine with being self sufficient. EMEA is content with sharing ideas and the North American LUAC is wanting to get more involved.
(Jump to slide 20) - Conclusions. Helping end users solve use and deployment issues will revitalize the rate and pace of Linux adoption. People are more interested in Linux in its ability to host. No one questions Linux anymore; it's based on the platforms readiness to host a business solution. The LUAC believes OSDL is uniquely positioned in helping users solve deployment issues. They realize OSDL can't solve everything. What success will the OSDL achieve? What are the ramifications? Also, how should they incorporate feedback from the LUAC? A key part is the revitalization plan - explains how to continue to deliver on areas. It tackles issues related to Linux as an end to end hosting platform, not just focusing on the kernel. Broker resources that solves the problem - basically put some light on the issue and what would be a successful outcome. The LUAC wants a to be in a position to function as an advisory council rather than a user group. A precaution is that in order for both sides to benefit they all need to come together at the solution. The LUAC believes OSDL can play a key role in collaboration for consensus to action. Slide 24 aligns OSDL activities with end user needs. It took about 3 meetings for the LUAC to come to a conclusion regarding their dissatisfaction. Some assumptions were being made that were incorrect about OSDL. They started analysis from the bottom up - how does OSDL operate, funding model, etc. They need to create a linkage and have the ability to drive outcome back to value as it's currently become rather vague. There is a desire to maintain a close linkage between the LUAC and OSDL work groups as the LUAC wants to see a more real time collaboration effort. The revitalization plan was put out there and some challenges that were outlined on slide 26. The last slide is a specific request of the OSDL board.
The TAB noted that they would be happy to get user feedback from the LUAC. There is overlapping membership from the TAB and Kernel Summit Program Committee who would also be interested in getting user feedback. The trouble is also getting users who are technically inclined enough to separate out distro issues and actual core kernel issues. There are multiple forms in which this can be channeled to the community either via KS or channeled to proper mailing lists. Marrying the right level of output and the communication vehicle would be a first step. Note that the input needs to be much more detailed and useful than just “you should work on virtualization”. It was however felt that this would also tie into the need for more funding and finding developers to contribute the code. Talk of the Fellowship Fund was brought up. OSDL to provide Tim with more information on the Fellowship Fund.
If the TAB has any further questions for Tim feel free to send him email (contact Leann for his address).
3. Review Fellowship Fund Request - TCP Testing
TAB felt this was a reasonable request to add to the pending queue. Motion was passed to add the TCP Testing Fellowship Fund Request to the pending queue. Will still need to acquire more donations to go forward with funding.
4. Hardware Lending Library - Adaptec boards
We have officially started this program. Let Tom know if there is hardware out there you want him to work on getting. Google is offering to set us up a hardware environment for the Adaptec boards but James hasn't secured their complete agreement. Whomever the board goes to needs to have an appropriate rig. Tom will wait to hear from James before dispersing the boards.
5. Technical Writer Status
Diane's changes look fine. OSDL will take care of the official posting externally. Was also asked if this should be sent over LKML as well? Job postings are typically frowned upon on LKML however this one sort of falls into the grey area. OSDL will post it first and if any TAB member thinks it's appropriate to push to LKML they can do so. Leann will send text file to James once we get the go ahead of posting. The TAB might want to forward to LWN as well. At some point, OSDL will make a press release, most likely once the hire is made.
6. TAB and community support goals for 2007 - 2009
OSDL has been working on strategic planning for the next few years. They want to be sure to keep the TAB involved. They want to open up the communication with what sorts of programs the TAB wants OSDL to incorporate into their strategic planning. The TAB felt the charter has a good initial list of topics. If the tab wants to refine or expand anything please send email to Tom or use the mailing list.eek October 1, 2006 - October 7, 2006