Download Full-Text PDF Cite this Publication

Text Only Version



1VIGNESH M C (Ph. no9019866126), 2 Dr MZ KURIAN

1,2SSIT, Tumkur, Karnataka


As of todays network architecture the complexity is very high and the usage of cores in each chip is very large and hence these factors are becoming an important issue to maintain the reliability factors. The agent based architecture proposed here will maintain the performance and reliability of the network based chips and this architecture proposed here will have higher tolerance towards the faulty links and router.

To utilize the best possible available resources the fault information is classified and enlightened among the neighboring units. The designed architecture has only small extra hardware architecture.

  1. The network on chip is found to be more reliable when used with Chip Multi-Processors (CMP) which have been designed and used to overcome the intrinsic design challenges in order to handle the large processing requirements. CMPs units may include hundreds of Intellectual Property (IP) cores, processing elements and embedded memory blocks to communicate with each other.

    When the number of nodes in the network increases the reliability, power consumption and thermal issues becomes an important factor. The reliability aspect of the network is the crucial consideration to be made. This aspect includes the proposed agent- based management structure and a routing method which is exploited to tolerate permanent faults in the nodes and links by using an agent. So the fault tolerances against permanent faults are also considered since a considerable amount of device failures may occur in both manufacturing and operational phases. To tolerate permanent faults many fault- tolerant routing algorithms have been designed so far. However, because of the size of CMPs and network architecture, we only consider distributed and scalable routing algorithms.

    The agent-based architecture consists of agents in a hierarchical structure, which is especially suitable for large CMPs.This structure where the agents in each level of hierarchy have the same tasks to gather, manage and distribute the fault information among the neighboring units.

    The Mesh network topological structure is used to evenly distribute the fault information and make sure the information is made to reach each and every node. So the appropriate information of the fault information will be sent to the direct and indirect neighbor nodes through the agents to be used in the routing process. This way, a scalable and fault-aware routing algorithm is achieved with higher performance compared to methods without agent-based management.

  2. Warsaw, in May 24-26, 2012.explained the complexity of evolving integrated circuits and the number of cores in each chip increase, reliability aspects are becoming an important issue in complex chip designs. The on-chip network architecture that incorporates a novel agent-based management method to enhance the reliability and performance of network-based Chip Multi-Processor (CMP) and System-on-Chip (SoC) are designed against faulty links and routers. In addition, to utilize the fault information required for the routing process in a scalable manner, we classify the fault information to be exploited in the proposed distributed and hierarchical management structure. The experimental results show that the proposed architecture incurs only a small hardware overhead [1].

    O. Cesariow, MODERN SYSTEM-ON-A-CHIP (SoC) design shows a clear trend toward integration of multiple processor cores. The SoC system driver section of the International Technology Roadmap for Semiconductors (http://public.itrs.net) predicts that the number of processor cores in a typical SOC will increase fourfold per technology node to match the corresponding applications processing demands. Typical multiprocessor SoC (MPSoC) applications, such as network processors, multimedia hubs, and baseband telecommunications circuits, have particularly tight time-to-market and performance constraints that demand a very efficient design cycle [2].

    C. Feng, Z. Lu, A. Jantsch, J. Li and M. Zhang, has proposed its known to everyone knows, developments in semiconductor technology have made very complex large scale System-on- Chip (SoCs) design available. Usually, in a System-on- Chip, interconnection between different Intellectual Property (IP) cores is achieved by means of shared bus architectures. However, nowadays each new SOC generation integrates more processing elements, more features and more new functionalities.

    With this increasing complexity, the system maybe needs

    extremely high capability in computation and communication. So, it is obvious that, use of shared bus architectures will affect performance of the overall system.

    In order to solve these problems, They have two different communication operations. Firstly, NIs can collect the data from all kinds of elements that are attached to them, packetize, add the header and tail information and send the processed packet into the attached router. Secondly, they receive the packets from the attached router. Network on Chip (NOC) is one solution for designing communication among components in the SOC circuits with several billion transistors that will reach the market in approximately 5-10years from now. One reason for a new communication strategy is that it is too costly to use one global synchronous clock in a circuit, since it will take several clock pulses just for a signal to travel across the chip [3].

    M. Valinataj, S. Mohammadi, and S. Safari Currently, the manufacturing process of complex systems such as network- based Systems-on-Chip incurs a considerable amount of failures. This paper presents a class of very low cost routing algorithms to increase the yield and to tolerate permanent faulty links in Networks-on-Chip. These new algorithms are fault tolerant through dynamic reconfiguration when the regular topology is altered by faulty links. Also, the proposed algorithms are the reconfigurable extensions of deterministic routing algorithms and their deadlock freeness is obtained by prohibiting a few turns. To demonstrate the effectiveness of the proposed routing algorithms, the performance, power consumption, and area overheads are evaluated through appropriate simulations and syntheses [4].

    M. Valinataj, S. Mohammadi, J. Plosila, P. Liljeberg, and H. Tenhunen, As everyone knows, developments in semiconductor technology have made very complex large scale System-on- Chip (SoCs) design available. Usually, in a System-on- Chip, interconnection between different Intellectual Property (IP) cores is achieved by means of shared bus architectures. However, nowadays each new SOC generation integrates more processing elements, more features and more new functionalities. With this increasing complexity, the system maybe needs extremely high capability in computation and communication. So, it is obvious that, use of shared bus architectures will affect performance of the overall system [5].

    P. Rantala, J. Isoaho and H. Tenhunen, Here a novel agent- based reconfiguring concept for futures network-on-chip (NoC) systems is introduced. The necessary properties to increase architecture level fault tolerance are introduced. The system control is modeled as multi-level agent hierarchy that is able to increase application fault-tolerance and performance with autonomous reactions of agents. The agent technology adds a system level intelligence level to te traditional NoC system design. The architecture and functions of this system are described on conceptual level. Communication and reconfiguring data flows are presented as study cases. Principles of reconfiguration of a NoC on faulty environment are demonstrated and simulated. Probability of reconfiguration success is measured with different latency requirements and amount of redundancy by

    Monte Carlo simulations. The effect of network topology in reconfiguration of a faulty mesh was also under research in the simulations [6].

    W. Yin on-chip point-to-point distributed interconnection networks or Networks on a Chip (NOCs) have been proposed. Unlike the shared bus architectures, the key communication method in NOC architecture is to implement interconnections of different IP cores using on chip packet-switched networks. NOCs are composed of three components: Routers (switches), Links and Network Interfaces (NI). Routers are the switching elements that are responsible for forwarding data packets from one router to another one. Links are the connection parts between different routers, and they are usually bidirectional. Network Interfaces are the wrapper between the router and processing element (PE) [7].

    L. Guang, B. Yang, J. Plosila, K. Latif, and H. Tenhunen and p.Rantala, j. isoaho proposed a case study for hierarchical agent monitoring design approach, which provides a high level abstraction for designing monitoring functions on massively parallel and distributed systems. The case study features hierarchical power monitoring on NoC platforms, where each level of agents perform specific monitoring operations based on their granularity.

    The monitoring hierarchy and operations are specified by a formal language for consistent and non-ambiguous system design. Various benchmarks are mapped onto NoCs, running with hierarchical power monitoring agents. Quantitative evaluations are performed in terms of energy efficiency, communication latency, and silicon overhead [8][9].

    1. Kohler, G. Schley, and M. Radetzki, Here The structural redundancy inherent to on-chip interconnection networks [networks on chip (NoC)] can be exploited by adaptive routing algorithms in order to provide connectivity even if network components are out of service due to faults, which will appear at an increasing rate with future chip technology nodes. This paper is based on a new, fine-grained functional fault model and a corresponding distributed fault diagnosis method that facilitate determining the fault status of individual NoC switches and their adjacent communication links. The network fault-tolerance assume switches to be either available or fully out of service, we present a novel adaptive routing algorithm that employs the remaining functionality of partly defective switches. [10].


    A clusture module receives the data from network interface units and using a receiver module and it is processed and classified according to its address specified. For this RAM and packet classification modules are used. In the cell agent all the information regarding the neighboring cell is made available and a test agent method makes sure of this operation and the control register is used to finalize the o/p.

    Over all modules:


    The clusture module consits of the Receiver, RAM, Packet classification units in it.


        Receiver is an electronic device that receives the information. The information received by the receiver will be in the form of images, text, sound etc.Receiver has been designed as shown below it consists of, Datain,clk,eop,error,rst,sop,valid,byte_count,data_out,error_out,val id_out.Here the input data has been given to the receiver module and data will be stored in the memory Architecture for supporting all kinds of bounded-modifiers.

        Features of reciever:Data_in receives the data from remote environment/applications. Availability of the clock increases the

        navigation of the function Clock has the range from 4MHZ to 25 MHZ. 76bytes of data out capacity.

        Clk, rst, valid, sop, error, eop are of single bit.


        Random access memory is a form of data storage. A random access device allows stored data to be accessed directly in any random order. In contrast other data storage media such as hard discs, CDs, DVDs and magnetic tapes read and write data in a predetermined order, because of mechanical design limitations. Therefore the time to access a given data location varies significantly depending on its physical location.

        Features of RAM: WE pin is used to select the RAM in read=0 or in write=1 mode.DI is a data-input pin and DO is the data output pin respectively with 76bytes of capacity.Whenever the reset is high the DI and DO operation is halted.Clk has speed ranging from 4 MHz to 25 MHzs.


        Routers classify packets to determine which flow they belong to and to decide what service they should receive. Classification may, in general, be based on an arbitrary number of fields in the packet header. Performing classification quickly on an arbitrary number of fields is known to be difficult, and has poor worst case performance.

        So the Packet classification module is one of the important blocks of the cluster module where the incoming packets are tested and routed to the intended blocks. Since the fault information is available on the cell agent the faulty paths are avoided here.

        Features of packet classification: LDA pin is used for load axess control.Byte count holds the no of counts or operation done for the i/p data.Ethernet frame header (ETH_header) is of 26bytes in length and is chosen based on IPV4.


    Clusture agent is the combination of receiver, RAM, and packet classification, which forms the clusture module. The cluster module is used as a processing element in a network architecture which is then used with the cell agents or router module to form the overall 4×4 architecture.


      A packeet generator is used to generate the random packet or the user defined data.



      It acts as an intermediate between client and server here we can consider the source and the Destination.


      This unit controls the overall behaviour of the processing element.This control resister performs various actions.


    The cell and clusture module combinly forms the on chip advanced network design using the method know as the agent based method.

    The complete 4×4 architecture can be seen below as shown where total of 16 modules can be developed.

    This 4×4 architecture is used in network to route the packets from one point to the destination point. Here the reliability is maintained to the best possible extent and the faulty units are completely avoided by information regarding the neighboring unit status.


    1. Clusture module.

    The result can be seen for various values of Data_in for the respective selection of VALID, LDA, ERROR and RST respectively.

    Condition 1: Only when RST is low further operation continuous.. Condition 2: Valid is an active LOW pin.

    Condition 3: EOP is an active high pin

    Condition 4: Byte count will keep the record of how many no of counts generated.

    Condition 5: when the above conditions are met the data given at the input will be available at the output.

    Practically the data from network is processed to the output i.e., it reaches the intended destination only.

    The result is developed based on the using Xilinx for writing code and simulating and modelsim a free student edition Output simultor.


The scalable agent-based management architecture for fault- tolerant NoC-based CMPs is proposed. To exploit the proposed agent-based architecture for the fault- tolerant routing process, different types of required fault information are classified to be distributed in the network. Each level of agent hierarchy manages and distributes a specific type of fault information. This way, a higher performance can be achieved in the networks of different sizes. Here, we will investigate the agents in different levels of hierarchy with a more efficient management process with higher architecture. In addition, the fault information will be classified and managed which will be useful for the routing process in the neighbor clusters.


  1. Warsaw Reliable On-Chip Network Design Using an Agent-based Management Method:"Mixed Design of Integrated Circuits and Systems",May 24-26, 2012, 19th International Conference, Poland,.

  2. O. Cesariow et al., Multiprocessor SoC platforms: a component-based design approach, IEEE Design and Test of Computers, vol. 19, no. 6, pp. 5263, 2002.

  3. C. Feng, Z. Lu, A. Jantsch, J. Li, and M. Zhang, FoN: Fault-on- Neighbor aware routing algorithm for Networks-on-Chip, Proc. 23th IEEE Int. System-on-Chip Conf. (SOCC), pp. 441446, 2010.

  4. M. Valinataj, S. Mohammadi, and S. Safari, Fault-aware and reconfigurable routing algorithms for Networks-on-Chip, IETE Journal of Research, vol. 57, no. 3, pp. 215223, 2011.

  5. M. Valinataj, S. Mohammadi, J. Plosila, P. Liljeberg, and H. Tenhunen, A reconfigurable and adaptive routing method for fault-tolerant mesh- based networks-on-chip, Elsevier, Int. J. Electronics and Communications (AEÜ), vol. 65, no. 7, pp. 630640, 2011.

  6. P. Rantala, J. Isoaho, and H. Tenhunen, Novel agent-based management for fault-tolerance in network-on-chip, Proc. 10th Euromicro Conf. on Digital System Design (DSD), pp. 551555, 2007.

  7. A. W. Yin et al, Hierarchical agent monitoring NoCs: a design methodology with scalability and variability, Proc. 26th NORCHIP Conf., pp. 202207, 2008.

  8. L. Guang, B. Yang, J. Plosila, K. Latif, and H. Tenhunen, Hierarchical power monitoring on NoC – a case study for hierarchical agent monitoring design approach, Proc. 28th NORCHIP Conf., 2010.

  9. L. Guang, E. Nigussie, P. Rantala, J. Isoaho, and H. Tenhunen, Hierarchical agent monitoring design approach towards self-aware parallel systems-on-chip, ACM Trans. on Embedded Computing Systems, vol. 9, no. 3, article 25, 2010.

  10. A. Kohler, G. Schley, and M. Radetzki, Fault tolerant network on chip switching with graceful performance degradation, IEEE Trans. on Computer-Aided Design of Integrated Circuits and Systems, vol. 29, no.6, 2010.

  11. M. Li, Q. Zeng, and W. Jone, DyXY- a proximity congestion-aware deadlock-free dynamic routing method for Network on Chip, Proc.43th Design Automation Conference (DAC), pp. 849852, 2006.

Leave a Reply

Your email address will not be published. Required fields are marked *