diff options
Diffstat (limited to 'scripts/PLAN')
| -rw-r--r-- | scripts/PLAN | 119 |
1 files changed, 18 insertions, 101 deletions
diff --git a/scripts/PLAN b/scripts/PLAN index a2958b8c..f0eb191e 100644 --- a/scripts/PLAN +++ b/scripts/PLAN @@ -1,85 +1,12 @@ $Id$ -Rough plan for initial prototype RE. - -1) Hack quick cert generator [done] - -2) Hack xml-parse-test into something that can: - - - Generate XML (just compose directly, probably) - - Encode XML in CMS using specified signing key and certs - - Write that to a destination file - - - Pick up signed CMS from from a file - - Check CMS signatures and extract XML - - Validate (xmllint --relaxng) the XML - - Decode (xml.sax) the XML and dump the result - - Log everything as we go. - data structures. - -Schema is RNC (RelaxNG Compact) syntax, xmllint wants RNG (RelaxNG -XML) syntax, so use /usr/ports/textproc/trang to convert: - - $ trang foo.rnc foo.rng - -3) Whack (2) into something that does this over minimal HTTP. - -4) [to be continued...] - - - -Fun with SQL. Tim has given me a nice SQL schema for the objects we -worked out in Prague (superset of left-right protocol). Still -debugging that design, but broad outlines are clear. - -Some question how to deal with in-memory data and SQL without going -mad. Am assuming RE is event driven. Current theory: pull in an -entire context (one row of the "self" table and everything that links -to that self_id), run this event with it, save out any changes before -this event terminates. Perhaps keep an LRU cache of the last few such -contexts in which we've run, to save loading every time (eg, receipt -of an answer to an up-down or left-right protocol query is a discrete -event, would be annoying to have to read everything back from SQL -again). - -If we can work out a tree that spans one such memory context (almost -there), we can just walk that tree to handle pretty much all SQL. - -Basic data structures we need for every object we're going to store in -SQL: - -- Object type to hold the object (duh). - -- List of children in the data structure tree. - -- SQL SELECT expression (including variable bindings!) to pull all - rows matching blah blah out of SQL. Tricky bit here is variable - bindings, as otherwise the obvious handling would be a simple - method. Various possible ways to hack this in Python, perhaps just - - lambda(x,y,z): """SQL blah blah""" % dict - - Or perhaps this meta-tree is itself built from a Python class - (probably should be) and these lambdas turn into methods of the - meta-tree. Or something. - -- SQL to INSERT ... - -- SQL to UPDATE ... - -- SQL to DROP ... - At this point it's starting to look like the pdu objects in rpki.left_right are a subclass of a more general class of objects that -live in SQL (duh, again). This makes sense as left-right protocol was +live in SQL (duh). This makes sense as left-right protocol was designed to let us frob a subset of the rpki database on the wire. -Need to check whether current rpki.left_right objects map cleanly to -data structures we're saving in SQL. They should, but check anyway. - Hmm, it's more complicated than that. Most of left_right maps well, but portions (eg, <list_resources/>) are just commands and do not. Mixin? We have a varient of this same problem with the mixins I'm @@ -114,13 +41,6 @@ saved when the current event exits. Haven't done anything about db.commit() and db.rollback() yet, for that matter haven't yet whacked MySQL to enable those features. -Hack du jour is rpki.left_right.data_obj.serve_copy_hook(). I've -implemented this for bsc objects, seems to work, haven't yet done (eg) -route_origin objects. - -left-right protocol needs boolean flags to reset array values like -bsc.signing_cert[]. Trivial hack, just haven't done it yet. - Access to object data attributes really ought to be through accessor methods so that the .set() method can set the sql_dirty flag automagically. Not done yet. @@ -134,23 +54,20 @@ magic names or config variables pointing at files. Makefile hack to put a Python wrapper around a .rng file, and automatically convert .rng to lxml internal form while we're at it? - - -More TODO: - -- create rpki-specific exception hierarchy (might be flat), convert - inappropriate uses of assert and standard exceptions to use - rpki-specific exceptions, then write exception handlers that - translate all of this into appropriate protocol error messages. - -- generalize serve_copy_hook() into serve_action_hook() and use it to - handle actions signaled by the various left-right boolean controls. - one fun test case for this should be bsc keypair generation. - - some of the defined actions may be things we can't implement yet due - to lack of a real timer system. eg, the various "do this right now" - actions are probably best handled as triggers, a la interrupt - handlers. well, maybe. irbe folks might prefer a version that - blocked response to the request until the whole cycle is done, once - we have an event system that can treat the query and response phases - of an rpc operation as two distinct events. +Have started on RPKI-specific exception hierarchy (well, it might be +flat) and conversion of inappropriate uses of assert and standard +exceptions to use rpki-specific exceptions. Top level of protocol +handlers needs to include exception catches that translate these +exceptions into appropriate protocol error messages. + +Generalize serve_copy_hook() into serve_action_hook() and use it to +handle actions signaled by the various left-right boolean controls. +One fun test case for this should be bsc keypair generation. + +Some of the defined actions may be things we can't implement yet due +to lack of a real timer system. Eg, the various "do this right now" +actions are probably best handled as triggers, a la interrupt +handlers. Well, maybe. IRBE folks might prefer a version that +blocked response to the request until the whole cycle is done, once we +have an event system that can treat the query and response phases of +an rpc operation as two distinct events. |