Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Tiny Microthreading Operating System: TinyOS Networked Sensor Characteristics Small physical size and low power consumption Concurrency-intensive operation Limited physical parallelism and controller hierarchy Diversity in Design and Usage Robust Operation Photograph and schematic for the Network Sensor The Solution: TinyOS A microthreaded OS that draws on previous work done for lightweight thread support, and efficient network interfaces Two level scheduling structure Long running tasks that can be interrupted by hardware events Small, tightly integrated design that allows crossover of software components into hardware TinyOS - Design Components Commands Events Scheduler Events Hardware components Structure of a Component Command Handlers Set of Tasks Event Handlers Frame (containing state information) TinyOS Component TinyOS Component Model Component has: Frame (storage) Tasks (computation) Command and Event Interface Messaging Component Internal Tasks Commands Internal State Events To facilitate modularity, each component declares the commands it uses and the events it signals. Statically allocated, fixed sized frames allow us to know the memory requirements of a component at compile time, and avoid the overhead associated with dynamic allocation. Tasks Perform the primary computation work Atomic with respect to other tasks, and run to completion, but can be preempted by events Allow the OS to allocate a single stack assigned to the currently executing task Call lower level commands Signal higher level events Schedule other tasks within a component Simulate concurrency using events Commands Non-blocking requests to lower level components Deposit request parameters into a component’s frame, and post a task for later execution Can also invoke lower level commands, but cannot block To avoid cycles, commands cannot signal events Return status to the caller Events Event handlers deal with hardware events (interrupts) directly or indirectly Deposit information into a frame Post tasks Signal higher level events Call lower level commands AM_MSG_SEND_DONE AM_MSG_REC AM_SEND_MSG AM_POWER AM_INIT TOS Component Messaging Component Commands AM_RX_PACKET _DONE Internal State AM_TX_PACKET _DONE AM_SUB_POWER AM_SUB_TX_PACKET AM_SUB_INIT Internal Tasks Events //AM.comp// TOS_MODULE AM; ACCEPTS{ char AM_SEND_MSG(char addr, char type, char* data); void AM_POWER(char mode); char AM_INIT(); }; SIGNALS{ char AM_MSG_REC(char type, char* data); char AM_MSG_SEND_DONE(char success); }; HANDLES{ char AM_TX_PACKET_DONE(char success); char AM_RX_PACKET_DONE(char* packet); }; USES{ char AM_SUB_TX_PACKET(char* data); void AM_SUB_POWER(char mode); char AM_SUB_INIT(); }; Putting It All Together The task scheduler is a simple FIFO scheduler. The scheduler puts the processor to sleep when the task queue is empty. Peripherals keep operating and can wake up the processor. Communication across components takes the form of a function call. This results in low overhead, and allows compile time type checking. Sample Application The sample application is consisting of a number of sensors distributed within a localized area Monitor temperature and light conditions Periodically transmit measurements to a base station Sample Application (cont) Sensors can forward data for other sensors that are out of range of the base station Dynamically determine the correct routing topology for the network Image courtesy Jason Hill et al Internal Component Graph Ad hoc Routing Application application Active Messages packet Radio Packet Serial Packet Temp Radio byte UART I2C SW byte HW Photo bit Clocks RFM Slide courtesy Jason Hill et al Ad hoc Routing Base station periodically broadcasts route updates Any sensors in range of this broadcast record the identity of the base station, and rebroadcast the update Each sensor remembers the first update received in an era, and uses the source of the update as the destination for routing data back to the base 2 station 1 3 Base 0 2 1 Image courtesy Jason Hill et al Resource requirements Review Component interface: commands accepts (implemented) commands uses events accepts (implemented) events uses Messaging Component Internal Tasks Internal State Component implementation functions that implement interface frame: internal state tasks: concurrency control Commands Events Programming Model comp3 comp1: C code Components .comp: specification .C: behaviour .desc: select and wire specification: comp2: .desc comp4 accepts commands uses commands signals events handles events application: .desc Programming Model 4 kinds of components: .desc only (ex: cnt_to_leds) .C and .comp (ex: CLOCK) .desc and .comp (ex: GENERIC_COMM) .C, .comp and .desc (ex: INT_TO_RFM) Programming Model (4th type) component := interface (.comp) + implementation (.c) + wiring (.desc) <CompName>.comp TOS_MODULE <CompName>; ACCEPTS { // command_signatures }; HANDLES { // event_signatures }; USES { // command_signatures }; SIGNALS { // event_signatures }; Programming Model <CompName>.c #include "tos.h" #include “<CompName>.h" #define TOS_FRAME_TYPE TOS_FRAME_BEGIN(< CompName >_frame) { // state declaration } TOS_FRAME_END(< CompName >_frame); char TOS_COMMAND(<command_name)(){ // command implementation } char TOS_EVENT(<event_name>)(){ // event implementation } Programming Model <CompName>.desc // Component Selection INCLUDE { MAIN; <CompName>; <Comp_I>; <Comp_J>; … }; // Wiring <CompName>.<command> <Comp_I>.<command> … <CompName>.<event> <Comp_J>.<event> … TOS 101: the Blink example example app that handles the clock events to update LEDs like a counter MAIN main_sub_init main_sub_start blink_init blink_start BLINK blink_sub_init clock_init CLOCK blink_clock_event blink_ledy_on blink_ledy_off clock_fire_event yellow_led_on yellow_led_off LED TOS 101: the Blink example blink.desc include modules { MAIN; BLINK; CLOCK; LEDS; }; BLINK:BLINK_INIT BLINK:BLINK_START BLINK:BLINK_LEDy_on BLINK:BLINK_LEDy_off BLINK:BLINK_LEDr_on BLINK:BLINK_LEDr_off BLINK:BLINK_LEDg_on BLINK:BLINK_LEDg_off BLINK:BLINK_SUB_INIT BLINK:BLINK_CLOCK_EVENT MAIN:MAIN_SUB_INIT MAIN:MAIN_SUB_START LEDS:YELLOW_LED_ON LEDS:YELLOW_LED_OFF LEDS:RED_LED_ON LEDS:RED_LED_OFF LEDS:GREEN_LED_ON LEDS:GREEN_LED_OFF CLOCK:CLOCK_INIT CLOCK:CLOCK_FIRE_EVENT TOS 101: the Blink example blink.comp TOS_MODULE BLINK; ACCEPTS{ //commands char BLINK_INIT(void); char BLINK_START(void); }; HANDLES{ // events void BLINK_CLOCK_EVENT(void); }; USES{ // commands char BLINK_SUB_INIT(char interval, char scale); char BLINK_LEDy_on(); char BLINK_LEDy_off(); char BLINK_LEDr_on(); char BLINK_LEDr_off(); char BLINK_LEDg_on(); char BLINK_LEDg_off(); }; SIGNALS{ //events }; TOS 101: the Blink example blink.c #include "tos.h" #include "BLINK.h" //Frame Declaration #define TOS_FRAME_TYPE BLINK_frame TOS_FRAME_BEGIN(BLINK_frame) { char state; } TOS_FRAME_END(BLINK_frame); /* BLINK_INIT: Clear all the LEDs and initialize state */ char TOS_COMMAND(BLINK_INIT)(){ TOS_CALL_COMMAND(BLINK_LEDr_off)(); TOS_CALL_COMMAND(BLINK_LEDy_off)(); TOS_CALL_COMMAND(BLINK_LEDg_off)(); VAR(state)=0; TOS_CALL_COMMAND(BLINK_SUB_INIT)(tick1ps); return 1; } TOS 101: the Blink example blink.c (cont.) /* BLINK_START: initialize clock component to generate periodic events. */ char TOS_COMMAND(BLINK_START)(){ return 1; } /* Clock Event Handler: Toggle the Red LED on each tick. */ void TOS_EVENT(BLINK_CLOCK_EVENT)(){ char state = VAR(state); if (state == 0) { VAR(state) = 1; TOS_CALL_COMMAND(BLINK_LEDr_on)(); } else { VAR(state) = 0; TOS_CALL_COMMAND(BLINK_LEDr_off)(); } } nesC Programming Language Goals: pragmatic low-level language for programming motes intermediate language for future high-level languages nesC Programming Language comp3 comp1: module Components: implementation - module: C behaviour - configuration: select and wire interfaces - comp2: configuration provides interface requires interface comp4 application: configuration nesC Programming Language 2 kinds of components: configuration: was .desc and .comp module: was .C and .comp nesC Blink example blink.td (configuration) configuration Blink { } implementation { uses Main, BlinkM, Clock, Leds; Main.SimpleInit -> BlinkM.SimpleInit; BlinkM.Clock -> Clock; BlinkM.Leds -> Leds; } nesC Blink example blinkM.td (module) module BlinkM { provides { interface SimpleInit; } requires { interface Clock; interface Leds; } } implementation { bool state; command result_t SimpleInit.init() { state=FALSE; return SUCCESS; } ... nesC Blink example blinkM.td (module) ... command result_t SimpleInit.start() { return call Clock.setRate(128, 6); } event result_t Clock.fire() { state = !state; if(state) call Leds.redOn(); else call Leds.redOff(); } } nesC Programming Language nesc1 takes a component and: checks nesC errors checks (most) C errors generates one C-file for the application avr-gcc compiles nesc1 output generated code has #line directives, so clean error messages nesC Programming Language Application C (TinyOS) nesC Savings Blink 1092 796 27% CntToLeds 1170 972 17% SenseToLeds 1636 1086 34% Oscilloscope 2792 2230 20% RfmToLeds 5226 4168 20% GenericBase 4544 4632 -2% CntToRfm 5220 4678 10% CntToLedsAndRfm 5544 4850 13% Chirp 5516 4948 10% EEPROMTest 7830 6220 21% nesC Programming Language nesC much nicer to use than TinyOS: produces smaller code than TinyOS get rid of the macros for frames, commands, events eliminate wiring error through type checking simplify wiring by using interfaces increase modularity by separation of interfaces and modules The Communication Subsystem Active Messages Active Messages : Motivation Legacy communication cannot be used : TCP/IP, sockets, routing protocols like OSPF Bandwidth intensive Centered on “stop and wait” semantics Need real time constraints and low processing overhead Active Messages Integrating communication and computation Matching communication primitives to hardware capabilities Provides a distributed eventing model where networked nodes send events to each other Closely fits the event-based model of TinyOS Active Messages Message contains : User-level handler to be invoked on arrival Data payload passed as argument Message handlers executed quickly to prevent network congestion and provide adequate performance Event-centric nature enables network communication to overlap with sensor-interaction Active Message + TinyOS = Tiny Active Messages Messaging is a component in TinyOS Tiny Active Messages Support for three basic primitives : Best effort message transmission Addressing Dispatch Three primitives necessary and sufficient Applications build additional functionality on top Tiny Active Messages - Details Packet format TinyDB-Design, Code and implementation TinyDB-Overview TinyDB is a query processing system for extracting information from a network of TinyOS sensors. TinyDB provides a simple, SQL-like interface. TinyDB provides a simple Java API for writing PC applications that query and extract data from the network TinyDB-Overview Contd.. Two Subsytems Sensor Network Software Sensor Catalog and Schema Manager Query Processor Memory Manager Network Topology Manager Java-based Client Interface Architecture TinyDB GUI TinyDB Client API JDBC PC side Mote side 0 0 2 1 4 TinyDB query processor 5 Sensor network 83 6 7 DBMS Data Model Entire sensor network as one single, infinitely-long logical table: sensors Columns consist of all the attributes defined in the network Typical attributes: Sensor readings Meta-data: node id, location, etc. Internal states: routing tree parent, timestamp, queue length, etc. Nodes return NULL for unknown attributes On server, all attributes are defined in catalog.xml Query Language (TinySQL) SELECT <aggregates>, <attributes> [FROM {sensors | <buffer>}] [WHERE <predicates>] [GROUP BY <exprs> [HAVING having-list]] [EPOCH DURATION integer] [INTO <buffer>] [TRIGGER ACTION <command>] Comparison with SQL Single table in FROM clause Only conjunctive arithmetic comparison predicates in WHERE and HAVING No subqueries No column alias in SELECT clause Arithmetic expressions limited to column op constant Only fundamental difference: EPOCH DURATION clause TinySQL Examples “Find the sensors in bright nests.” Sensors 1 SELECT nodeid, nestNo, light FROM sensors WHERE light > 400 EPOCH DURATION 1s Epoch Nodeid nestNo Light 0 1 17 455 0 2 25 389 1 1 17 422 1 2 25 405 TinySQL Examples (cont.) 2 SELECT AVG(sound) FROM sensors EPOCH DURATION 10s “Count the number occupied nests in each loud region of the island.” Epoch 3 SELECT region, CNT(occupied) region CNT(…) AVG(…) 0 North 3 360 FROM sensors 0 South 3 520 GROUP BY region 1 North 3 370 HAVING AVG(sound) > 200 1 South 3 520 AVG(sound) EPOCH DURATION 10s Regions w/ AVG(sound) > 200 Event-based Queries ON event SELECT … Run query only when interesting events happens Event examples Button pushed Message arrival Bird enters nest Analogous to triggers but events are user-defined Query over Stored Data Named buffers in Flash memory Store query results in buffers Query over named buffers Analogous to materialized views Example: CREATE BUFFER name SIZE x (field1 type1, field2 type2, …) SELECT a1, a2 FROM sensors EPOCH DURATION d INTO name SELECT field1, field2, … FROM name EPOCH DURATION d Tree-based Routing Tree-based routing Used in: Query delivery Data collection In-network aggregation Relationship to indexing? Q:SELECT … A Q R:{…} Q R:{…} B Q R:{…}Q Q D R:{…}Q C Q Q R:{…} Q Q Q F E Q Power Management Approach Coarse-grained app-controlled communication scheduling Mote ID 1 … zzz … Epoch (10s -100s of seconds) … zzz … 2 3 4 5 time 2-4s Waking Period Time Synchronization All messages include a 5 byte time stamp indicating system time in ms Synchronize (e.g. set system time to timestamp) with Any message from parent Any new query message (even if not from parent) Punt on multiple queries All nodes agree that the waking period begins when (system time % epoch dur = 0) And lasts for WAKING_PERIOD ms Adjustment of clock happens by changing duration of sleep cycle, not wake cycle. TinyDB and the Java API TinyDBNetwork sendQuery() injects query into network abortQuery() stops a running query addResultListener() adds a ResultListener that is invoked for every QueryResult received removeResultListener() SensorQueryer translateQuery() converts TinySQL string into TinyDBQuery object Key difference from JDBC: push vs. pull TinyDB and Java API Contd.. TinyDBQuery a list of attributes to select a list of expressions over those attributes, where an expression is a filter that discards values that do not match a boolean expression an aggregate that combines local values with values from neighbors, and optionally includes a GROUP BY column. an SQL string that should correspond to the expressions listed above. QueryResult A complete result tuple, or A partial aggregate result, call mergeQueryResult() to combine partial results TinyDB and Java API Contd.. AggOp SelOp Catalog CommandMsgs TinyDB Demo Application TinyDBMain Opens AM (Active message) connection to serial port(“COM1”) and uses it to initialize a Tiny\DBNetwork object. Allocates GUI objects CmdFrame and QueryFrame CmdFrame - Sending TinyDB commands into the network MainFrame - Main GUI for building queries QueryField - Routines for handling attributes in the query builder ResultFrame Plot Topology Result Graph Magnet Frame Inside TinyDB Sensor Catalog and Schema Manager Query Processor Memory Manager Network Topology Manager Component Diagram Inside TinyDB Schema-capabilities of motes in the system as a single virtual Database “table” Tables-typed attributes , commands that can be run within query executor. Query processing-sensor readings from each mote is placed in “tuples” (passed between motes for multihop routing and/or aggregation, or might be passed to the font end code) Attr and command components Inside TinyDB Contd.. Components: TinyDBAttr TinyDBCommand Tuple QueryResult TinyDB Query Processing operators TupleRouter - Heart of TinyDB System The TupleRouter component contains three main execution paths: Handling of new query messages Result computation and propagation (each time a clock event goes off) Subtree result message handling TinyDB Query Processing operators Contd.. Network.queryMsg event- new queries Query message-part of a query: either a single field (attribute) to retrieve, a single selection predicate to apply, or a single aggregation function. parseQuery() – Compact representation of query. Allocation of space given a parsed query setSampleRate() Deliver the tuples that were completed Decrement the counter for all queries Fetch data fields Route filled-in tuples to query operators. Network.dataMsg event –Neighbour result arrival SelOperator AggOperator TinyDB MultiHop Routing Modular interface -Network.nc. The royting layers must provide the following methods. command QueryResultPtr getDataPayLoad(TOS MsgPtr msg) command TinyDBError sendDataMessage(TOS MsgPtr msg) command QueryMessagePtr getQueryPayLoad(TOS MsgPtr msg) command TinyDBError sendQueryMessage(TOS MsgPtr msg) event result t sendQueryDone(TOS MsgPtr msg, result t success): event result t sendDataDone(TOS MsgPtr msg, result t success): event result t dataSub(QueryResultPtr qresMsg): event result t querySub(QueryMessagePtr qMsg): event result t snoopedSub(QueryResultPtr qresMsg, bool isFromParent, uint16 t senderid) Inside TinyDB SELECT T:1, AVG: 225 AVG(temp) Queries Results T:2, AVG: 250 WHERE light > 400 Multihop Network Query Processor Aggavg(temp) ~10,000 Lines Embedded C Code Filter light > 400 got(‘temp’) ~5,000 LinesSamples (PC-Side) Java get (‘temp’) Tables Schema ~3200 Bytes RAM (w/ 768 byte heap) getTempFunc(…) TinyOS code ~58 kB compiled (3x larger than 2nd largest TinyOS Program) TinyDB Extending TinyDB Why extending TinyDB? New sensors attributes New control/actuation commands New data processing logic aggregates New events Analogous to concepts in object-relational databases TinySchema Collection of TinyOS components that manages a small repository of named attributes, commands and events. Attribute-similar to column in traditional database. Command-stored procedure in a traditional database system. Actuation commands - physical actions on motes Tuning commands - adjust internal parameters Event-capture asynchronous events in WSN, e.g, detection of a bird. TinySchema Contd.. Attribute interfaces Stdcontrol (initialization) AttrRegister (create new non-constant attribute) AttrRegisterConst (create new constant attribute) AttrUse (discover and use) Command Interfaces Stdcontrol (initialization) CommandRegister (create new command) CommandUse (discover and use) Event Interfaces StdControl (initialization) EventRegister (create new event) EventUse (discover, signal) Adding Attributes Types of attributes Sensor attributes: raw or processed sensor readings Introspective attributes: values from internal software or hardware states e.g parent node in routing tree, voltage, ram usage, etc. Constant attributes: constant values that can be statically or dynamically assigned to a mote, e.g., nodeid, location, etc. Adding Attributes (cont) Interfaces provided by Attr component StdControl: init, start, stop AttrRegister command registerAttr(name, type, len) event getAttr(name, resultBuf, errorPtr) event setAttr(name, val) command getAttrDone(name, resultBuf, error) AttrUse command startAttr(attr) event startAttrDone(attr) command getAttrValue(name, resultBuf, errorPtr) event getAttrDone(name, resultBuf, error) command setAttrValue(name, val) Adding Attributes (cont) Steps to adding attributes to TinyDB 1) 2) 3) 4) Create attribute nesC components Wire new attribute components to TinyDBAttr configuration Reprogram TinyDB motes Add new attribute entries to catalog.xml Constant attributes can be added on the fly through TinyDB GUI TinyDB Aggregation Framework Adding Aggregates (cont) Step 2: add entry to catalog.xml <aggregate> <name>AVG</name> <id>5</id> <temporal>false</temporal> <readerClass>net.tinyos.tinydb.AverageClass</readerClass> </aggregate> Step 3 (optional): implement reader class in Java a reader class interprets and finalizes aggregate state received from the mote network, returns final result as a string for display. TinySec TinySec Architectural Features Single shared global cryptographic key Link layer encryption and integrity protection transparent to applications Cryptography based on a block cipher K K K TinySec Summary Security properties Access control (Only nodes that posses the shared key can participate in the network) Integrity (a message should only be accepted if it was not altered in transit) Confidentiality (the content of a message shoulb not be infered by unauthorized parties) TinySec (The fact) Integration OS: TinyOS 1.1.0 Processors: Mica, Mica2, Mica2Dot using Atmel Processors Radio: RFM TR1000 and Chipcon CC1000 SIM: TOSSIM simulator Implementation 3000 lines of NesC code RAM: 455 bytes (not an issue for applications, can be reduced to 256 bytes) MEM: 7000 bytes of program space Real time: Two priority TinyOS scheduling process (cryptographic computations must be completed by the time the radio finishes sending the start symbol) Usage Build: maintains a key file and uses a key from the file, includes the key at compile time. Application: “make TINYSEC=true …” to enable TinySec-Auth. TinySec (Interfaces) App App GenericComm SecureGenericComm Radio Radio Making deployment easy: plug-n-play crypto + link-layer security TinySec Interfaces TinySec TinySecM: bridges radio stack and crypto libraries BlockCipher 3 Implementations: RC5M, SkipJackM, IdentityCipher BlockCipherMode CBCModeM: handles cipher text stealing No length expansion when encrypting data MAC CBCMACM: computes message authentication code using CBC TinySec (Components) Interface: TinySec TinySecM Radio Stack [MicaHighSpeedRadioM/ CC1000RadioIntM] CBC-MACM CBC-ModeM Interface: SkipJackM BlockCipher BlockCipherInfo Use a block cipher for both encryption & authentication Skipjack is good for 8-bit devices; low RAM overhead TOSSIM MOTIVATION Embedded nature of sensor networks makes controlled experiments difficult Development is complicated by motes' small form factor and limited accessibility in the field Inspecting the internal state of programs on many remote nodes is laborious Inspection that disturbs the reactive nature of a mote (ie. Breakpoint) can invalidate the observed behaviour TOSSIM TOSSIM: A discrete event simulator for TinyOS sensor networks sensor networks Provide high fidelity (reliability/accuracy) simulation of TinyOS applications Debug, test and analyze algorithms in a controlled and repeatable manner Key Requirements: Scalability, Completeness, Fidelity, Bridging TOSSIM Bridging Implementations, not just algorithms Completeness Cover as many system interactions as possible Fidelity Capture these interactions at a very fine grain Scalability Examine behavior in dense or large networks (Largest sensor network ever deployed: 850 nodes) TOSSIM Scales to thousands of nodes Simulates the TinyOS network at the bit level Compiles directly from TinyOS code ('make pc') Flexibility to replace parts of application component graph (ie. Packet level radio component) Useful for prototyping and development (debugging, breakpoints etc.) IMPLEMENTATION NOTES Compiler support – nesC (ncc) Execution model – interrupts and events Hardware - abstracts each resource as a component (ie. ADC, Clock, EEPROM, boot seq, radio stack) Radio Models – can choose accuracy & complexity of model (allows for error-free transmission) Communication services – mechanisms to allow PC apps to drive, monitor etc. TOSSIM What TOSSIM Doesn't do: Times interrupts, not execution time Does not model radio propogation Does not model power draw or energy consumption (but you can extend to track consumption (power/time stats) Interrupts are nonpreemptive TinyViz COMPONENTS Communication subsystem Event bus Synchronization, information passing Plug-ins Drawing, mote options Subscribe to events Send commands Maintain state GUI Drawing, user interaction TinyViz ARCHITECTURE Visualizing Simulation Actuating Simulation This is the last slide ! Time to wake up http://www.i2r.astar.edu.sg/icsd/SecureSensor/Document.html