First cut at cleaning 0.1.1.x TODO

svn:r4879
This commit is contained in:
Nick Mathewson 2005-08-27 03:20:51 +00:00
parent 2e7de08fbf
commit c6c403e156

171
doc/TODO
View File

@ -14,9 +14,8 @@ PHOBOS - phobos claims
Non-Coding, Soon:
N - contact umass folks
N - Packaging logic and HOWTO for controller libs
N - Mention controller libs someplace.
- FAQ entry: why gnutls is bad/not good for tor
D FAQ entry: why gnutls is bad/not good for tor
P - flesh out the rest of the section 6 of the faq
P - gather pointers to livecd distros that include tor
- put the logo on the website, in source form, so people can put it on
@ -26,7 +25,7 @@ P - gather pointers to livecd distros that include tor
* clean up the places where our docs are redundant (or worse, obsolete in
one file and correct elsewhere). agl has a start on a global
list-of-tor-docs.
P - update window's docs to clarify which versions of windows, and why a
P - update windows docs to clarify which versions of windows, and why a
DOS window, how it's used, for the less technical users
NR- write a spec appendix for 'being nice with tor'
- tor-in-the-media page
@ -34,16 +33,13 @@ NR- write a spec appendix for 'being nice with tor'
tor-0.1.0.7.rc
- Remove need for HACKING file.
For 0.1.0.x:
. Memory use on Linux: what's happening?
- Is it threading? (Maybe, maybe not)
- Is it the buf_shrink bug? (Quite possibly)
- Instrument the 0.1.1 code to figure out where our memory is going;
apply the results. (all platforms?)
for 0.1.1.x:
R - are dirservers auto-verifying duplicate nicknames?
N . Additional controller features
- Find a way to make event info more extensible
- change circuit status events to give more details, like purpose,
whether they're internal, etc.
. Expose more information via getinfo:
@ -54,116 +50,92 @@ N . Additional controller features
download directories/network-status, and a way to force a download.
- It would be nice to request address lookups from the controller
without using SOCKS.
N . helper nodes (Choose N nodes randomly; if a node dies (goes down for a
long time), replace it. Store nodes on disk.
o Implement (basic case)
o Implement (persistence)
o Document
. Test, debug
- On sighup, if usehelpernodes changed to 1, use new circs.
. Helper nodes
. More testing and debugging
- On sighup, if usehelpernodes changed to 1, use new circuits?
- If your helper nodes are unavailable, don't abandon them unless
other nodes *are* reachable.
R - If you think an OR conn is open but you can never establish a circuit
to it, reconsider whether it's actually open.
- switch accountingmax to count total in+out, not either in or
out. it's easy to move in this direction (not risky), but hard to
back, out if we decide we prefer it the way it already is. hm.
. Come up with a coherent strategy for bandwidth buckets and TLS. (The
logic for reading from TLS sockets is likely to overrun the bandwidth
buckets under heavy load. (Really, the logic was never right in the
first place.) Also, we should audit all users of get_pending_bytes().)
- Make it harder to circumvent bandwidth caps: look at number of bytes
sent across sockets, not number sent inside TLS stream.
. Handle rendezvousing with unverified nodes.
o Specify: Stick rendezvous point's address and port in INTRODUCE cell.
o Handle new format.
o Support to extend circuit/target circuit to a chosen combination of
addr/port/ID/onionkey
o Parse new format
o Generate new format (#ifdef out the logic to generate it for now)
o Specify: make service descriptors contain onion key and identity.
o Implement new service desc format
o Think: are we okay with the partitioning? (Yes. It's a simple
migration issue.)
o Implement new directory code
o Implement new server code (Don't enable till directory code is deployed)
o Implement new client code (Don't enable till directory code is deployed)
o Look for v1 descriptor if available, else look for v0 descriptor.
o Use new INTRODUCE protocol if allowed.
N . Verify that new code works.
- Enable the new code
- christian grothoff's attack of infinite-length circuit.
- Miscellaneous cleanups
- switch accountingmax to count total in+out, not either in or
out. it's easy to move in this direction (not risky), but hard to
back, out if we decide we prefer it the way it already is. hm.
. Come up with a coherent strategy for bandwidth buckets and TLS. (The
logic for reading from TLS sockets is likely to overrun the bandwidth
buckets under heavy load. (Really, the logic was never right in the
first place.) Also, we should audit all users of get_pending_bytes().)
- Make it harder to circumvent bandwidth caps: look at number of bytes
sent across sockets, not number sent inside TLS stream.
R - remove the warnings from rendezvous stuff that shouldn't be warnings.
N . Handle rendezvousing with unverified nodes.
o Implement everything
. Enable the new code
. Verify that new code works.
- Christian Grothoff's attack of infinite-length circuit.
the solution is to have a separate 'extend-data' cell type
which is used for the first N data cells, and only
extend-data cells can be extend requests.
- Specify, including thought about
- Implement
N - Destroy and truncated cells should have reasons.
N - Add private:* alias in exit policies to make it easier to ban all the
fiddly little 192.168.foo addresses.
(AGL had a patch; consider applying it.)
- recommended-versions for client / server ?
N - warn if listening for SOCKS on public IP.
- cpu fixes:
- see if we should make use of truncate to retry
o hardware accelerator support (configure engines.)
- hardware accelerator support (use instead of aes.c when reasonable)
R - kill dns workers more slowly
R - remove the warnings from rendezvous stuff that shouldn't be warnings.
- continue decentralizing the directory
o Specify and design all of the below before implementing any.
- Figure out what to do about hidden service descriptors.
X have two router descriptor formats
R . dirservers verify reachability claims
o basic reachability testing, influencing network-status list.
R - rate-limiting the reporting of trouble servers
R - check reachability as soon as you hear about a new server
- find 10 dirservers. (what are criteria to be a dirserver?)
- some back-out mechanism?
. Directory changes
o recommended-versions for client / server ?
- Some back-out mechanism for auto-approval
- dirservers have blacklist of IPs they hate
- a way of rolling back approvals to before a timestamp
- have new people be in limbo and need to demonstrate usefulness
before we approve them
- other?
N . Authoritative dirservers publish very compressed network-status objects.
o Generate format
o Publish it
N . Everyone downloads network-status objects
- From all directories, round-robin
- Cache them, reload on restart
o Serve cached directories
- If DirPort, act as a cache.
N - Directories expose individual descriptors
o By server ID
o By 'all'
- By 'if-newer-than' (Does the spec require this??)
- Support compression.
o Expose "own most recent descriptor".
N - Alice acts on network-status objects, downloading descriptors as needed.
o Servers publish new descriptors when:
o options change
o when 12-24 hours have passed
o when uptime is reset
o When bandwidth changes a lot.
- alices avoid duplicate class C nodes.
o everybody with a dirport will give you his descriptor.
- config option, on by default, to cache all descriptors.
- Compress router desc sets before transmitting them
M Analyze how bad the partitioning is or isn't.
- Naming:
- Specify and design all of the below before implementing any.
- some dirservers announce that they manage bindings (a flag in
router-status).
- other dirservers mention a binding if there is no conflict for
that binding among the dirservers that manage it.
no conflict == any of them bind it and no disagreement.
- alice can specify a nickname and it will record that name in her
datadir along with the key *if* it is bound. otherwise her specifying
will fail (loudly we hope).
- thus when a binding vanishes (e.g. conflict) alice will keep using
the one she meant.
- if the binding changes keys, the entry in her datadir will silently
get corrected.
R . Dirservers verify reachability claims
o basic reachability testing, influencing network-status list.
R - rate-limiting the reporting of trouble servers
R - check reachability as soon as you hear about a new server
- Decentralization
- Figure out what to do about hidden service descriptors.
- find 10 dirservers.
- (what are criteria to be a dirserver?)
N . Dirservers publish compressed network-status objects.
- Support several-at-once
N . Everyone downloads network-status objects
- From all directories, round-robin
- Cache them, reload on restart
o Serve cached directories
N . Directories expose individual descriptors
X By 'if-newer-than' (Does the spec require this??)
- Support compression.
N - Alice acts on network-status objects
- Alice downloads descriptors as needed.
- Alice sets descriptor status from networks-status
- Security
- Alices avoid duplicate class C nodes.
- Analyze how bad the partitioning is or isn't.
N - Naming:
- Separate naming from validation in authdirs.
- Clients choose names based on network-status options.
- Names are remembered in client status.
- packaging and ui stuff:
. multiple sample torrc files
- uninstallers
@ -175,15 +147,18 @@ N - Alice acts on network-status objects, downloading descriptors as needed.
N - Vet all pending installer patches
- Win32 installer plus privoxy, sockscap/freecap, etc.
- Vet win32 systray helper code
o Make logs go into platform default locations.
o OSX
X Windows. (?)
Reach (deferrable) items for 0.1.1.x:
- Start using create-fast cells as clients
o Let more config options (e.g. ORPort) change dynamically.
- start handling server descriptors without a socksport?
. Research memory use on Linux: what's happening?
- Is it threading? (Maybe, maybe not)
- Is it the buf_shrink bug? (Quite possibly)
- Instrument the 0.1.1 code to figure out where our memory is going;
apply the results. (all platforms?)
For 0.1.1.x, if we can figure out how:
- rewrite how libevent does select() on win32 so it's not so very slow.
o enclaves (at least preliminary)