Engineering Swarm- Services CS

Transcription

1 Engineering Swarm- Services CS Spring 2014 John Wawrzynek Adam Wolisz Technische Universitat Berlin UC Berkeley EECS Adjunct Lecture 01, Introduction 1 CS294-97, UC Berkeley Spring 14

2 Beacons Transmitted periodically by the AP to locate and identify its BSS (heartbeat of the BSS) Uses same channel access method as all other packets, i.e., is ordinary packet Header + BODY + CRC Beacon Interval Timestamp SSID (note: BSSID is part of the MAC address field) Supported rates Parameter: like channel number... Capabilities (like Support of Privacy) Traffic Indication Map (for whom packets are waiting) NOTE: An SSID is the Name of a Network BSSIDs Identify Access Points and Their Clients

3 Scanning A station must scan to find a BSS (independent or infrastructure) to join Passive scanning - just listening for traffic Active scanning - soliciting for responses Choice of scanning modes is an implementation decision Passive scanning Station listens for beacon and probe response frames An AP usually issues 10 beacons per second, i.e., the scan duration can be up to 100 ms per channel Changes to a different channel if nothing is heard or to learn of other BSSs Reordering the channel search can fasten-up the BSS discovery,,

4 Active Scanning Station sends a probe request on a channel Individual request (with specified SSID, for THE one) or broadcast request- any network should respond! Last node to transmit a beacon will respond Another station in an independent BSS The AP (always!!) in an infrastructure BSS Station may change channels if it does not receive a response or to find another BSSs Minimizes time to find a BSS at the expense of transmitting probes The time to receive a probe response is usually 10-20ms

5 Joining a BSS A station might discover MULTIPLE BSS! Joining a BSS is a local decision, i.e., there is no notice sent externally that a station has joined a particular BSS Station must set Time Synchronization Function (TSF) timer (coming up), which also synchronizes for FHSS PHY functions PHY parameters as advertised BSSID WEP and high rate (802.11b) capabilities Beacon interval for synchronization (coming up) Delivery Traffic Indication Map (DTIM) period for power management (coming up)

6 Synchronization Synchronization necessary for time-related functions for the PHY layer Provided by a timer synchronization function (TSF) Each station maintains a 64-bit counter running at a 1 MHz rate Timer reset to 0 when a station begins operation Updated by information in beacon frames

7 Synchronization using a Beacon (infrastructure) AP is responsible for sending beacon frames AP tries to send beacons at a Target Beacon Transmission Time (TBTT) TBTT is advertised in the beacon May be delayed since beacons must contend for the medium using DCF May be lost due to corruption or collision TBTT AP B B B B Medium busy busy busy busy t value of the timestamp B beacon frame accumulate time AP sends current counter value in beacon Station updates its counter to this value (plus processing time)

8 Synchronization using a Beacon (ad hoc) Synchronization must be fully distributed One station starts the BSS Resets its TSF counter to 0 Sends a beacon with a TBTT value This establishes a beacon process During each beacon interval, every station prepares to send a beacon Waits for a random time after the TBTT to begin transmission If a beacon is first received from another station, its beacon transmission is cancelled TBTT Station 1 B 1 B 1 Station 2 Medium B 2 B 2 busy busy busy busy t value of the timestamp B beacon frame random delay

9 Synchronization using a Beacon (ad-hoc) A station updates its counter only if the received value is greater than its own count Discarded otherwise This tends to propagate the time of the station with the fastest clock through the IBSS

10 Association with active scanning

11 Association reassociation... Only in infrastructure mode Station sends association request to an AP, including Capabilities, e.g., data rate, contention-free support, WEP support Request for contention-free service- if applicable... Length of time it may be in a low-power mode (see later) Access point responds with success or failure status Acceptance based on policies not specified in the standard, e.g. loading, capabilities, etc. AP provides an association ID (AID) to the station

12 Association (2) Once associated, the AP is responsible for delivering frames to and from the station From the station To other stations in the BSS To other stations in the ESS To other stations via the portal To the station While it is in the AP s BSS While it is elsewhere in the ESS Delivery to a station outside the BSS requires the distribution service (DS) to know the location of the station

13 Association (3) A station may need to reassociate itself with another AP via a re-association request All information in association request Plus BSSID of last AP Allows new AP to retrieve frames held by old AP Allows old AP to delete association with the station Station may only be associated with one BSS within a single ESS Theoretically could be multi-homed in multiple ESSs

14 Internet End-to-End View [Stoica] Process A sends a packet to process B Proc. A (port 10) Internet Proc. B (port 7) IP Address: A four-part number used by Network Layer to route a packet from one computer to another

15 Process Address to receive messages, process must have identifier identifier includes both IP address and port numbers associated with process on host. example port numbers: HTTP server: 80 Mail server: 25 to send HTTP message to gaia.cs.umass.edu web server: IP address: Port number: 80

16 Streams of Bits/bytes can be transmitted: so what? How do we know what is the INFORMATION inside?

17 Simple example Representation of base types floating point: IEEE 754 versus non-standard integer: big-endian versus little-endian (e.g., 34,677,374) (2) (17) (34) (126) Big- endian (126) (34) (17) (2) Little- endian Low address High address on a 680x0 CPU, the 32 bit integer number 255 is stored as: but an Intel 80x86-CPU stores this as:

18 Taxonomy Data types base types (e.g., ints, floats); must convert flat types (e.g., structures, arrays); must pack complex types (e.g., pointers); Conversion Strategy canonical intermediate form receiver-makes-right (an N x N solution)

19 Data Conversion Two different types of rules are needed: Abstract syntax: a station must define what datatypes are to be transmitted Transfer syntax: it must be defined how these datatypes are transmitted, i.e. which representation has to be used. Tagged versus untagged data type = INT len = 4 value =

20 Abstract Syntax Notation.1 - ASN.1 Each transmitted data value belongs to an associated data type. For the lower layers of the OSI-RM, only a fixed set of data types is needed (frame formats), for applications with their complex data types ASN.1 provides rules for the definition and usage of data types. ASN.1 distinguishes between a data type (as the set of all possible values of this type) and values of this type (e.g. 1 is a value of data type Integer). Basic ideas of ASN.1: Every data type has a globally unique name (type identifier) Every data type is stored in a library with its name and a description of its structure (written in ASN.1) A value is transmitted with its type identifier and some additional information (e.g. length of a string).

21 Service Service: Any act or performance that one party can offer to another that is essentially intangible and does not result in the ownership of anything. Its production may or may not be tied to a physical product. D. Jobber, Principles and Practice of Marketing Focus is on the output, the result of the service NOT the means to achieve it

22 Service Service: Any act or performance that one party can offer to another that is essentially intangible and does not result in the ownership of anything. Its production may or may not be tied to a physical product. D. Jobber, Principles and Practice of Marketing Focus is on the output, the result of the service NOT the means to achieve it

23 Remote Procedure Calls (RPC) Remote Procedural Calls are the preferred tool to implement the client-server model. In classical procedure calls the code of the procedure is located on the same computer (in the same address space) as the calling program, in an RPC the code is located on another computer. One major design goal of an RPC system is transparency: ideally the caller should not know if the callee is located locally or remotely. So in RPC we have to consider the following topics: Parameter handling and marshalling Semantics Addressing An RPC system is attractive for the users because automatic support for the conversion from local to remote procedural call can be supported (see below).

24 Marshaling [Karl, Paderborn] Marshalling: taking parameters/results of a procedure call and prepare them for transmission over a network To ensure, e.g., transparency between different hardware, operating systems, programming languages Handled by client stub & server stub/skeleton

25 Finding an RPC server (Addressing) A client can use fixed, hard coded addresses for finding the appropriate server station. This approach is simple but not flexible. A dynamic binding approach can be used: A server stub transmits at its initialization a message containing its name (procedure name), its version number, its address and a unique (within the server station) identification to a special station, the bindery station, which maintains a database of all available services. A client stub, if operating the first time, queries the bindery station for an appropriate server providing the requested service (i.e. service name, version number). If no server exists, the client stub fails. Otherwise, the bindery returns the address and the unique identification to the client stub.

26 Local vs. Remote Procedural Call