4 // Fraunhofer Institute for Open Communication Systems (FOKUS)
6 // The contents of this file are subject to the Fraunhofer FOKUS Public License
7 // Version 1.0 (the "License"); you may not use this file except in compliance
8 // with the License. You may obtain a copy of the License at
9 // http://senf.berlios.de/license.html
11 // The Fraunhofer FOKUS Public License Version 1.0 is based on,
12 // but modifies the Mozilla Public License Version 1.1.
13 // See the full license text for the amendments.
15 // Software distributed under the License is distributed on an "AS IS" basis,
16 // WITHOUT WARRANTY OF ANY KIND, either express or implied. See the License
17 // for the specific language governing rights and limitations under the License.
19 // The Original Code is Fraunhofer FOKUS code.
21 // The Initial Developer of the Original Code is Fraunhofer-Gesellschaft e.V.
22 // (registered association), Hansastraße 27 c, 80686 Munich, Germany.
25 // Stefan Bund <g0dil@berlios.de>
28 /** \mainpage libPPI : The Packet Processing Infrastructure
30 The PPI provides an infrastructure to create packet oriented network processing applications. A
31 PPI application is built by combining processing modules in a very flexible manner.
33 \image html scenario.png Target Scenario
35 The PPI concept is built around some key concepts
37 \li The PPI is based on processing \ref ppi_packets. It does not handle stream oriented
39 \li The PPI is built around reusable \ref ppi_modules. Each module is completely independent.
40 \li Each module has an arbitrary number of \ref ppi_connectors, inputs and outputs.
41 \li The modules are connected to each other using flexible \ref ppi_connections.
42 \li Data flow throughout the network is governed via flexible automatic or manual \ref
43 ppi_throttling (throttle notifications).
44 \li Modules may register additional external \ref ppi_events (file descriptor events or timers).
46 The PPI thereby builds on the facilities provided by the other components of the SENF
47 framework. The target scenario above depicts a diffserv capable UDLR/ULE router including
48 performance optimizations for TCP traffic (PEP). This router is built by combining several
51 \see \ref ppi_overview \n
52 <a href="../../../../Examples/RateStuffer/doc/html/index.html">PPI Example Application:
54 \ref senf::ppi::module "Modules" \n
55 \ref senf::ppi::connector "Connectors and Jacks" \n
59 /** \page ppi_overview PPI Overview and Concepts
63 \section ppi_design Design considerations
65 The PPI interface is designed to be as simple as possible. It provides sane defaults for all
66 configurable parameters to simplify getting started. It also automates all resource
67 management. The throttling infrastructure handles blocking conditions (like input exhaustion)
70 \section ppi_packets Packets
72 The PPI processes packets and uses the <a
73 href="@TOPDIR@/senf/Packets/doc/html/index.html">Packet library</a> to handle them. All packets
74 are internally passed around as generic \ref senf::Packet references, however connectors may
75 optionally be defined as sending or receiving packets of a specific type only.
77 \section ppi_modules Modules
79 A module is represented by a class derived from senf::ppi::module::Module. Each module has
82 \li It may have any number of \ref ppi_connectors (inputs and outputs)
83 \li Each module declares flow information which details the route packets take within the
84 module. This information does not define how the information is processed, it only tells,
85 where data arriving on some input will be directed at (\ref
86 senf::ppi::module::Module::route())
87 \li The module might take additional parameters.
88 \li The module might also register additional \ref ppi_events.
90 Generally, modules are divided into several categories:
92 \li \ref io_modules receive external data or forward packets out of the PPI
93 \li \ref routing_modules forward packets within the network
94 \li \ref adapter_modules are used to connect incompatible connectors to each other
95 \li Application modules are modules implemented to perform an applications function
97 Of these modules, normally only the application modules need to be implemented since the library
98 provides an extensive set of reusable modules.
100 The following example module declares three \ref ppi_connectors "Connectors": \c payload,
101 \c stuffing and \c output. These connectors are defined as \e public data members so they
102 can be accessed from the outside. This is important as we will see below.
106 : public senf::ppi::module::Module
108 SENF_PPI_MODULE(RateStuffer);
110 senf::ppi::IntervalTimer timer_;
113 senf::ppi::connector::ActiveInput<> payload;
114 senf::ppi::connector::ActiveInput<> stuffing;
115 senf::ppi::connector::ActiveOutput<> output;
117 RateStuffer(unsigned packetsPerSecond)
118 : timer_(1000u, packetsPerSecond)
120 route(payload, output);
121 route(stuffing, output);
123 registerEvent( timer_, &RateStuffer::tick );
137 The constructor will declare flow information using senf::ppi::module::Module::route(). Then the
138 module registers an interval timer which will fire <tt>packetsPerSecond</tt> times every
139 <tt>1000</tt> milliseconds.
141 The module processing is very simple: Whenever a timer tick arrives a packet is sent. If the \c
142 payload input is ready (see \ref ppi_throttling), a payload packet is sent, otherwise a stuffing
143 packet is sent. The module will therefore provide a constant stream of packets at a fixed rate
144 on \c output (see the
145 <a href="@TOPDIR@/Examples/RateStuffer/doc/html/index.html">RateStuffer</a> example application
146 for a slightly different approach)
148 An example module to generate the stuffing packets could be
151 class CopyPacketGenerator
152 : public senf::ppi::module::Module
154 SENF_PPI_MODULE(CopyPacketGenerator);
156 senf::ppi::connector::PassiveOutput<> output;
158 CopyPacketGenerator(Packet template)
159 : template_ (template)
162 output.onRequest(&CopyPacketGenerator::makePacket);
170 output(template_.clone());
175 This module just produces a copy of a given packet whenever output is requested.
177 \see senf::ppi::module::Module
179 \section ppi_connectors Connectors
181 The input and output attachment points of a module are called connectors. Each connector may be
182 active or passive. This gives us 4 types of connectors:
184 \li senf::ppi::connector::ActiveInput
185 \li senf::ppi::connector::ActiveOutput
186 \li senf::ppi::connector::PassiveInput
187 \li senf::ppi::connector::PassiveOutput
189 An \e active connector (input or output) is <em>activated by the module</em> to send data or to
190 poll for available packets. This means, the modules processing routine will call the connector
191 without being signaled by the framework to read the connector. It just actively fetches a
194 A \e passive connector is <em>signaled by the framework</em> to fetch data from the module or to
195 pass data into the module. The module must register a callback which will be called, whenever a
196 packet is requested from the module or whenever a new packet is made available for the module to
199 To send or receive a packet (either actively or passively) the module just calls the
200 connector. It is permissible to generate or process multiple packets in one iteration. However,
201 you must ensure yourself that enough packets are available to be read if more than one packet
202 shall be read. It is also permissible to not handle a packet at all even if signaled to do
203 so. The packet will automatically be queued.
205 To provide this flexibility, all input connectors incorporate a packet queue. This queue is
206 exposed to the module and allows the module to optionally process packets in batches.
208 Connectors take an optional template argument which allows to specify the type of packet this
209 connector sends or received. This template arguments defaults to \ref senf::Packet.
211 \see \ref senf::ppi::connector
213 \section ppi_connections Connections
215 \image html ratestuffer.png Simple RateStuffer
217 To make use of the modules, they have to be instantiated and connections have to be created
218 between its connectors. It is possible to connect any pair of input/output connectors as long as
219 one of them is active and the other is passive.
221 It is possible to connect two active or passive connectors with each other using a special
222 adaptor module (senf::ppi::module::PassiveQueue or senf::ppi::module::ActiveFeeder
225 Additionally, the connectors must be type-compatible: Either one (or both) of the connectors
226 must be untyped (they accept arbitrary senf::Packet's, the optional tempalte argument is empty),
227 or they both accept the same type of packet. This check is performed at runtime.
229 To complete our simplified example: Lets connect senf::ppi::module::ActiveSocketReader and
230 senf::ppi::module::PassiveSocketWriter to our example module:
233 RateStuffer rateStuffer (10);
235 senf::Packet stuffingPacket = senf::DataPacket::create(...);
236 CopyPacketGenerator generator (stuffingPacket);
238 senf::UDPv4ClientSocketHandle inputSocket (1111);
239 senf::ppi::module::ActiveSocketSource<> udpInput (inputSocket);
241 senf::UDPv4ClientSocketHandle outputSocket ("2.3.4.5:2222");
242 senf::ppi::module::PassiveSocketSink<> udpOutput (outputSocket);
244 senf::ppi::module::PassiveQueue adaptor;
246 senf::ppi::connect(udpInput, adaptor);
247 senf::ppi::connect(adaptor, rateStuffer.payload);
248 adaptor.qdisc(ThresholdQueueing(10,5));
249 senf::ppi::connect(generator, rateStuffer.stuffing);
250 senf::ppi::connect(rateStuffer, udpOutput);
255 This application will read udp-packets coming in on port 1111 and will forward
256 them to port 2222 on host 2.3.4.5 with a fixed rate of 10 packets / second.
258 We start out by instantiating the necessary modules. Then the connections between these modules
259 are set up by successively connecting each output connector to an input connector. As can be
260 seen, the name of the connector can be left of if it is named \c output or \c input
263 The buffering on the udpInput <-> rateStuffer adaptor is changed so the queue will begin to
264 throttle only if more than 10 packets are in the queue. The connection will be unthrottled as
265 soon as there are no more than 5 packets left in the queue (see \ref ppi_throttling).
267 \section ppi_throttling Throttling
269 Throttling and throttle notifications at it's base is about handling blocking conditions. The
270 most straight forward blocking condition is that of a file descriptor not being available for
271 reading resp. writing. Other blocking conditions can arise for example when a queue fills up or
272 if a module for some application specific reason does not want to handle packets for a period of
275 All this is handled using throttle notifications. We need throttle notifications so a passive
276 connector can tell it's connected peer that it cannot service further requests until an
277 unthrottle notification is sent. This tells us, that from the view of someone implementing a
278 module, throttle notifications will always be received on active connectors and be sent on
281 This tells us, that the direction of control flow (the throttle notifications) is from passive
282 to active connectors and does \e not depend on the direction of data flow (which flows from
283 output to input connector). Thinking about this, this makes sense: The module with the active
284 connector is the one initiating the data processing (after all, it is the \e active part) and
285 needs to be told not to request or send packets on it's connector since the connected passive
286 peer cannot handle the request.
288 So if a passive connector cannot handle requests, the connector must be \e throttled. Throttling
289 the connector will forward a throttle notification to its peer. The peer then handles the
290 throttling notification.
292 There are two ways, throttle notifications can be handled: By automatic throttling or by
293 registering callbacks. The default is <em>automatic throttling</em>.
295 <em>Automatic throttling</em> is based on the routing information available to the module. Every
296 notification received is forwarded within the module along all known routes from active to
297 passive connectors (routes which connect to active or passive connectors are absolutely valid,
298 they just are not \e forwarding routes, they are ignored by the throttle
299 notifications). Together with automatic event throttling (see \ref ppi_events), this is all that
300 is normally needed to handle throttle notifications: By forwarding the notifications we ensure,
301 that a module's passive connectors will only be signaled when it's corresponding active
302 connectors are not throttled (as defined by the routing information). The module is therefore
303 not called until the connector(s) are untrhottled.
305 <em>Throttle callbacks</em> can optionaly be registerd (with automatic throttling enabled or
306 disabled, see \ref senf::ppi::connector::ActiveConnector) to be called when a throttle
307 notification is received. The callback may then handle the notification however it sees fit, for
308 example by manually throttling some passive connector (see \ref
309 senf::ppi::connector::PassiveConnector).
311 To enable/disable automatic throttling, the \ref senf::ppi::module::Module::route() command
312 returns a reference to a \ref senf::ppi::Route instance. If this route is \e forwarding route,
313 (that is, of the connectors is passive and the other is active), the return value will be
314 derived from \ref senf::ppi::ForwardingRoute which provides members to control the throttle
315 notification forwarding.
318 senf::ppi::module::Module \n
321 \section ppi_events Events
323 Modules may register additional events. These external events are very important since they
324 drive the PPI framework. Events are like external calls into the module network which are sent
325 whenever some event happens. Some possible events are
326 \li timer events (senf::ppi::IntervalTimer)
327 \li read or write events on some file descriptor (senf::ppi::IOEvent)
328 \li internal events (senf::ppi::IdleEvent)
330 The PPI really is not concerned, how the events are called and what information is needed to
331 perform the call. This is handled by the <a
332 href="@TOPDIR@/senf/Scheduler/doc/html/index.html">Scheduler</a>, which is wrapped by the event
335 All events are derived from senf::ppi::EventDescriptor. The base class allows to enable and
336 disable the event. Each type of event will take descriptor specific constructor arguments to
337 describe the event to be generated. Events are declared as (private) data members of the
338 module and are then registered using senf::ppi::module::Module::registerEvent().
340 Each event when signaled is described by an instance of the descriptor specific \e
341 descriptorType \c ::Event class. This instance will hold the event specific information (like
342 scheduled time of the event, file handle state and so on). This information is passed to the
345 Additionaly, events are valid routing targets. This feature allows events to be disabled and
346 enabled by throtling notifications. For the sake of routing, an event may be used like an active
347 input or output. Iit is \e active from the PPI's point of view since it is signaled from the
348 outside and not by some module. It may be either input or output depending on the operation the
351 If we take into account event routing, we can extend the \c RateStuffer constructor accordingly:
354 RateStuffer(unsigned packetsPerSecond)
355 : timer_(1000u, packetsPerSecond)
357 route(payload, output);
358 route(stuffing, output);
359 route(timer_, output); // (*)
361 registerEvent( timer_, &RateStuffer::tick );
365 We have added the marked route call. This way, the \c timer_ will receive throttling
366 notifications from the output: Whenever the output is throttled, the event will be disabled
367 until the output is unthrottled again.
369 \see senf::ppi::EventDescriptor
371 \section ppi_run Running the network
373 After the network has been set up, senf::ppi::run() is called to execute it. This call will only
374 return after all data has been processed. The PPI knows this, when no events are enabled any
375 more. Without events, nothing will happen any more since it is the events which drive the
376 PPI. Therefore the PPI surmises, that all data has been processed and returns from
379 This works very well with automatic throttling. When no data is available to be processed any
380 more and no more data can be expected to arrive (for Example since data has been read from a
381 file which is now exhausted) all events will be disabled automatically via trhottle
382 notifications and so signal that any processing should stop.
384 \section ppi_flows Information Flow
386 The above description conceptually introduces three different flow levels:
388 \li The <em>data flow</em> is, where the packets are flowing. This flow always goes from output
390 \li The <em>execution flow</em> describes the flow of execution from one module to another. This
391 flow always proceeds from active to passive connector.
392 \li The <em>control flow</em> is the flow of throttling notifications. This flow always proceeds
393 \e opposite to the execution flow, from passive to active connector.
395 This is the outside view, from without any module. These flows are set up using
396 senf::ppi::connect() statements.
398 Within a module, the different flow levels are defined differently depending on the type of
401 \li The <em>data flow</em> is defined by how data is processed. The different event and
402 connector callbacks will pass packets around and thereby define the data flow
403 \li Likewise, the <em>execution flow</em> is defined parallel to the data flow (however possible
404 in opposite direction) by how the handler of one connector calls other connectors.
405 \li The <em>control flow</em> is set up using senf::ppi::Module::route statements (as long as
406 automatic throttling is used. Manual throttling defines the control flow within the
407 respective callbacks).
409 In nearly all cases, these flows will be parallel. Therefore it makes sense to define the \c
410 route statement as defining the 'conceptual data flow' since this is also how control messages
411 should flow (sans the direction, which is defined by the connectors active/passive property).
413 \see \ref ppi_implementation
416 /** \page ppi_implementation Implementation Notes
418 \section processing Data Processing
420 The processing in the PPI is driven by events. Without events <em>nothing will happen</em>. When
421 an event is generated, the called module will probably call one of it's active connectors.
423 Calling an active connector will directly call the handler registered at the connected passive
424 connector. This way the call and data are handed across the connections until an I/O module will
425 finally handle the request (by not calling any other connectors).
427 Throttling is handled in the same way: Throttling a passive connector will call a corresponding
428 (internal) method of the connected active connector. This method will call registered handlers
429 and will analyze the routing information of the module for other (passive) connectors to call
430 and throttle. This will again create a call chain which terminates at the I/O modules. An event
431 which is called to be throttled will disable the event temporarily. Unthrottling works in the
434 This simple structure is complicated by the existence of the input queues. This affects both
435 data forwarding and throttling:
436 \li A data request will only be forwarded, if no data is available in the queue
437 \li The connection will only be throttled when the queue is empty
438 \li Handlers of passive input connectors must be called repeatedly until either the queue is
439 empty or the handler does not take any packets from the queue
442 \section ppi_logistics Managing the Data Structures
444 The PPI itself is a singleton. This simplifies many of the interfaces (We do not need to pass
445 the PPI instance). Should it be necessary to have several PPI systems working in parallel
446 (either by registering all events with the same event handler or by utilizing multiple threads),
447 we can still extend the API by adding an optional PPI instance argument.
449 Every module manages a collection of all it's connectors and every connector has a reference to
450 it's containing module. In addition, every connector maintains a collection of all it's routing
453 All this data is initialized via the routing statements. This is, why \e every connector must
454 appear in at least one routing statement: These statements will as a side effect initialize the
455 connector with it's containing module.
457 Since all access to the PPI via the module is via it's base class, unbound member function
458 pointers can be provided as handler arguments: They will automatically be bound to the current
459 instance. This simplifies the PPI usage considerably. The same is true for the connectors: Since
460 they know the containing module, they can explicitly bind unbound member function pointers to
463 \section ppi_random_notes Random implementation notes
465 Generation of throttle notifications: Backward throttling notifications are automatically
466 generated (if this is not disabled) whenever the input queue is non-empty \e after the event
467 handler has finished processing. Forward throttling notifications are not generated
468 automatically within the connector. However, the Passive-Passive adaptor will generate
469 Forward-throttling notifications whenever the input queue is empty.
471 \section ppi_classdiagram Class Diagram
473 \diaimage classes.dia
480 // c-file-style: "senf"
481 // indent-tabs-mode: nil
482 // ispell-local-dictionary: "american"