Internet網(wǎng)絡(luò)技術(shù)課程論文

1、Internet網(wǎng)絡(luò)技術(shù)課程論文The Research on Computer Communication NetworksReal-time Communication院系：物理電氣信息學(xué)院 班級：08級電子班姓名：李偉 學(xué)號：12008243872The Research on Computer Communication NetworksReal-time CommunicationAbstract: Computer communication network has been a popular technology in the information society. Howe

2、ver, the problem of network communications real-time performance and networks quality of service (QOS) is the key technology of research by human all along. To solve this problem, the characteristic of network communication was analyzed firstly, and then computer network communications time constrai

3、nt condition was analyzed on the system point of view. Finally, the Intranet and Internets real-time performance was analyzed and evaluated respectively. Some conclusions are provided for practicality.Key Words: Network ; Ethernet; real-time communication; 1. Introduction of real-time communicationT

4、he Modern computer technology and the modern communication technology is combined to computer network technology, and more than two independent operating systems to connect to each other through the media to share a collection of hardware and software components of the computer network resources.Wit

5、h the extensive application of computer networks, more and more people put forward many requirements including a dedicated multimedia communications and control of the industrial measurement real-time communication. It is include a comprehensive Quality of Service (QOS). Traditional computer network

6、s also use packet switching technology is design for Non-real-time data communications and using TCP/IP protocol is mainly to optimize the overall network data throughput and data communications to provide best-effort service. But todays multimedia communications and industrial measurement control r

7、eal-time network communication not only includes traditional text data, but also includes such as audio, video and other information. In the WAN, Internet use the protocols TCP/IP achieve a global internetworking. In the LAN, some of the main network operating system also provides the protocol conve

8、rsion gateway software, but these Internet access has some limitations and generally do not provide heterogeneous network directly between the computer access real-time communication. However, in many application fields often the most stress is the real-time communication.In order to Make better use

9、 of computer networks and share information resources for data communications, that interconnection of heterogeneous networks and real-time communication is of great significance.2. Real-time communications feature The Ethernet of CSMA / CD algorithm before sending the data for each site will monito

10、r whether there are other sites in the channel to send data. If not, then send the data. However, due to the presence of propagation delay and the channel using this approach to pre-test can not be absolute to avoid conflict. Therefore, CSMA / CD requires sending side of the site while monitor. If i

11、t is listen to the conflict both of side first stop sending, and then send a length of strong interference signals k for other web sites to inform, and wait a random time after the process is repeated. When the information starts sending, it is need to check whether a collision, such as collision, t

12、hen the exit re-issued. Ethernet have a problem of uncertainty and poor real-time. Uncertainty refers to any node in the network for any load condition within a specified time can be the opportunity to send data packets and any node can not be exclusive transmission medium. But the main of real-time

13、 is through the time to reflect. When the real-time data and non real-time data transmission simultaneously on Ethernet. As the real-time data and non-real-time data in the source node of the competition, it is from other nodes with real-time and non real-time data of the collision. Real-time data w

14、ill be experienced unpredictable delay even the case for a long time unable to. It is may be let exclusive bus node failure caused by other nodes in the transmission failure and real-time network communication can not be meet the requirements. 3. The working principle of real-time communication Real

15、-time communication between computers through network transfer large amounts of data. Characterized by a high-speed and real-time that is requires high data transfer rate and low delay. Generally speaking, high speed depends on network hardware, and the real-time is closely related with data communi

16、cation. Real-time communication more mature approach is to use end-to-end communication, also known as peer communication. It works: From the OSI reference model perspective of the application layer, it can be known as network is composed of clients and server. However, transport layer and network l

17、ayer from the perspective of the client and server is not essentially different. They are connected to a machine on the network are available to replace the network address or name. They are in the peer network entity. Communication between them is that called end-to-end communication. The original

18、source machine and the final letter of a physical machine to the other entities placed to communicate directly between with each other as a direct line. Regardless of the transmission process to go through a number of intermediate nodes.End to end communication has three characteristics, it can grea

19、t degree of assurance of real-time communication. (1)When source and places have a connection, once the data is sent from a source who knows Sink will be able to receive data; (2) End to end data transmission without intermediate nodes storage of other entities forwarding. It can improve the transmi

20、ssion efficiency and give small delay; (3) the end-to-end high level software do not have a function of store and forward. It is directly dependent on the underlying software to store and forward and it will be more simply.4. On-Line Hard Real-Time Scheduler ARTCS For traditional network, maximizing

21、 the throughput or minimizing the average message delay is the most important performance criteria. In the hard real-time domain, however, concern focuses on satisfying the time constraints of individual message. The on-line hard real-time scheduler ARTCS presents an approach to absolutely guarantee

22、 deadlines of periodic and aperiodic 非周期性 messages in centralized-scheduing hard real-time communication systems on field bus domain. In the centralized-scheduling hard real-time communication system, the access to the bus is controlled by a node on which the bus scheduling table runs. The bus sched

23、uling table deals with the message as the basic scheduling unit. Each scheduling instruction in the bus table indicates the execution information of every message including the source and destination address, the beginning time and the maximal delay of the message execution, and the other control in

24、formation. By orderly executing the instructions of this bus table, not only the sending time of a nodes data, but also the channel time used by a node will be controlled. Consequently the hard real-time property of communication is guaranteed. Using traditional scheduling algorithms directly in a r

25、eal-time communication system, the transmitting message may be interrupted and has ti retransmit all over again, which leads to waste of the bandwidth and more swapping overhead. The PCP avoids messages transmission interruption, while a higher-priority message might be blocked until a lower-priorit

26、y message completion. Sometimes blocking of the higher-priority message even once may be too long to meet its deadline. On the other hand, when the PCP is applied to give the schedulable analysis for a set of many messages, calculating the blocked times of every messages is complex and not well suit

27、able in the practical industrial computing system. Nevertheless, the Sr can reduce the overhead of judging schedulability 調(diào)度 since it transforms the distance constrains into harmonics. Therefore, the ARTCS uses Sr to judge the schedulability of the message set. In the process of ARTCS, it applies th

28、e non-blockage and non-interruption approach to adapt the network communication. The ARTCS, based on RM, Sr and PS scheduling algorithms, includes three stages: building message set, judging schedulability and generating bus table. In the following section, the three stages are presented in detail.4

29、.1 Collecting Information and Building Message Set ARTCS collects all information relating to message transmission when the network system is initialized, then it creates corresponding original message set. The message set model is consistent with the practical application demand, either periodic or

30、 distance constrained message model. Let M=(Ci,Di)l=i=n,D1=D2=Dn be the message set, where Ci is the worse-case execution time and Di is the deadline (or distance constraint) of Mi. Meantime the aperiodic messages are translated into periodic messages by taking the minimal arrival intervals as their

31、 periods. 4.2 Judging Scheduability As we pointed out above, it is difficult to calculate blocked time of every message in practice when the PCP is used. In order to reduce the computational complexity, the ARTCS judges the schedulability of the message set based on the Sr. For the periodic message

32、model, the ARTCS regards the deadline Di as the distance constraint, and converts the periodic message model into the distance constrained message model. The judging schedulability stage of ARTCS includes two steps: base judgment and emulation judgment. 4.2.1 ARTCS Schedulability Base Judgment It is

33、 supposed that message execution can be preempted during the course of ARTCS schedulability base judgment. According to Sr scheduling algorithm, base judgment works as follows. Let r =Dor r = respectively (D is the maximal distance constraint among the messages of the set M), specialize the set D= (

34、D, D, D) with respect to (r), then D is transformed into D=(bb,b). Let P(t)I ).Therefore, for the transformed set M, the least common multiple (LCM) of all distance constraints is equal to b, so we only consider the schedulability over the duration of period b. The method of the emulation judgment i

35、s as follows; firstly, divide the binto k segments, each of them is b; secondly, schedule the highest priority message that is ready among the message set, and the highest priority message will be scheduled only if the message execution can be completed in the current segment, otherwise it will be r

36、emoved to next segment; thirdly, whenever all message can be scheduled and satisfy their distance constraints during the period b, the set M can be scheduled successfully on condition that message execution isnt interrupted and blocked. This process called emulation judgment is showed as follow: /*M

37、=M(C,D) , 1=i=n is a set of messages that will be judged, with D=b,b,b, and p(M)= =C) 10) R;=R-C;11) if (next=n) return success/*emulation judgment is success*/12) next:= next+1;13) 14) 15) Record(K)/* register the time allocation of kth segment */16) if(Isint(logk)then 17) P:=logk+1;18) if (! Sched

38、ule (P) return failure /*emulation judgment is failure*/19) 20) In above processes, according to distance constraints, the messages is sorted into all of kinds , and each message execution interval is equal to its period. If the pth kind period length b(except b) is divided into tow parts averagely,

39、 for all messages whose periods are lower than b, scheduling them in each segment (segment length is b) of the later part is the same as the previous part entirely. Moreover, when the system allocates time for a certain segment (such as the kth segment) in the later part, it will firstly schedule th

40、e message whose periods are lower than b consistent with the corresponding reference segment (the jth segment is the reference segment of the kth segment) in the previous part. If the messages whose periods are lower than bhave been scheduled wholly from the 1st to the jth segments, given d is the t

41、otal time that is allocated to all messages whose periods are lower than bin the reference segment, let residual time R be equal to bminus d, then system decides the allocation of residual time according to whether R is bigger than the execution time C next of the first message that has never been s

42、cheduled. This process protects message execution from being interrupted and preempted. Line 57 represent the judgment whether the messages form 1st to jth segments whose periods are lower than b are scheduled wholly; line 814 show how to allocated the residual time of the kth segment. Whenever each

43、 kind period is end, the system will check whether the messages belong to this period have been scheduled wholly. If it is true, the process will turn to the next segment, otherwise the schedulability emulation judgment is failure. If the system has schedulted all messages within b, the schedulabili

44、ty emulation judgment is proved successful.5. Conclusions In this article over the computer network communications systems of the background, we generally understand the basic principles of real-time communications and have a rough concept of real-time communication. By using the improved pinwheel scheduling algorithm to judgment the schedulability, the ARCTS is able to no