« back to all changes in this revision
Viewing changes to doc/developers/performance-roadmap-rationale.txt
- browse files at revision 2485
- compare with another revision
- compare with revision 2490.2.32
- download diff
- download tarball
- view history from revision 2485
- Committer: Canonical.com Patch Queue Manager
- Date: 2007-05-10 05:55:01 UTC
- mfrom: (2481.1.5 roadmap)
- Revision ID: pqm@pqm.ubuntu.com-20070510055501-w262sk5hl33vmd19
(robertc) Move HACKING to doc/developers/HACKING and start on performance analysis documentation. (Robert Collins)
- files added:
- doc/developers
- files renamed:
- HACKING => doc/developers/HACKING
- files modified:
- .bzrignore
added
removed
doc/developers/performance-roadmap-rationale.txt
1
What should be in the roadmap?
2
------------------------------
3
4
A good roadmap provides a place for contributors to look for tasks, it
5
provides users with a sense of when we will fix things that are
6
affecting them, and it also allows us all to agree about where we are
7
headed. So the roadmap should contain enough things to let all this
8
happen.
9
10
I think that it needs to contain the analysis work which is required, a
11
list of the use cases to be optimised, the disk changes required, and
12
the broad sense of the api changes required. It also needs to list the
13
inter-dependencies between these things: we should aim for a large
14
surface area of 'ready to be worked on' items, that makes it easy to
15
improve performance without having to work in lockstep with other
16
developers.
17
18
Clearly the analysis step is an immediate bottleneck - we cannot tell if
19
an optimisation for use case A is a pessimism for use case B until we
20
have analysed both A and B. I propose that we complete the analysis of
21
say a dozen core use cases end to end during the upcoming sprint in
22
London. We should then be able to fork() for much of the detailed design
23
work and regroup with disk and api changes shortly thereafter.
24
25
I suspect that clarity of layering will make a big difference to
26
developer parallelism, so another proposal I have is for us to look at
27
the APIs for Branch and Repository in London in the light of what we
28
have learnt over the last years.
29
30
What should the final system look like, how is it different to what we have today?
31
----------------------------------------------------------------------------------
32
33
One of the things I like the most about bzr is its rich library API, and
34
I've heard this from numerous other folk. So anything that will remove
35
that should be considered a last resort.
36
37
Similarly our relatively excellent cross platform support is critical
38
for projects that are themselves cross platform, and thats a
39
considerable number these days.
40
41
And of course, our focus on doing the right thing is what differentiates
42
us from some of the other VCS's, so we should be focusing on doing the
43
right thing quickly :).
44
45
What we have today though has grown organically in response to us
46
identifying bottlenecks over several iterations of back end storage,
47
branch metadata and the local tree representation. I think we are
48
largely past that and able to describe the ideal characteristics of the
49
major actors in the system - primarily Tree, Branch, Repository - based
50
on what we have learnt.
51
52
What use cases should be covered?
53
---------------------------------
54
55
My list of use cases is probably not complete - its just the ones I
56
happen to see a lot :). I think each should be analysed comprehensively
57
so we dont need to say 'push over the network' - its implied in the
58
scaling analysis that both semantic and file operation latency will be
59
considered.
60
61
These use cases are ordered by roughly the ease of benchmarking, and the
62
frequency of use. This ordering is so that when people are comparing bzr
63
they are going to get use cases we have optimised; and so that as we
64
speed things up our existing users will have the things they do the most
65
optimised.
66
67
* status tree
68
* status subtree
69
* commit
70
* commit to a bound branch
71
* incremental push/pull
72
* log
73
* log path
74
* add
75
* initial push or pull [both to a new repo and an existing repo with
76
different data in it]
77
* diff tree
78
* diff subtree
79
80
* revert tree
81
* revert subtree
82
* merge from a branch
83
* merge from a bundle
84
* annotate
85
* create a bundle against a branch
86
* uncommit
87
* missing
88
* update
89
* cbranch
90
91
how is development on the roadmap coordinated?
92
----------------------------------------------
93
94
I think we should hold regular get-togethers (on IRC) to coordinate on
95
our progress, because this is a big task and its a lot easier to start
96
helping out some area which is having trouble if we have kept in contact
97
about each areas progress. This might be weekly or fortnightly or some
98
such.
99
100
we need a shared space to record the results of the analysis and the
101
roadmap as we go forward. Given that we'll need to update these as new
102
features are considered, I propose that we use doc/design as a working
103
space, and as we analyse use cases we include them in there - including
104
the normal review process for each patch. We also need documentation
105
about doing performance tuning - not the minutiae, though that is
106
needed, but about how to effective choose things to optimise which will
107
give the best return on time spent - that is what the roadmap should
108
help with, but this looks to be a large project and an overview will be
109
of great assistance I think. We want to help everyone that wishes to
110
contribute to performance to do so effectively.
111
112
Finally, its important to note that coding is not the only contribution
113
- testing, giving feedback on current performance, helping with the
114
analysis are all extremely important tasks too and we probably want to
115
have clear markers of where that should be done to encourage such
116
contributions.
Loggerhead is a web-based interface for Breezy
Version: 2.0.1