Você está na página 1de 4

JOURNAL OF COMPUTING, VOLUME 4, ISSUE 4, APRIL 2012, ISSN 2151-9617 https://sites.google.com/site/journalofcomputing WWW.JOURNALOFCOMPUTING.

ORG

95

Class of Service (CoS) Translations in Butterfly MPLS Network-on-Chip


Mohammad Reza Nouri Rad
Department of Computer Engineering Islamic Azad University KhorramAbad Branch, Iran

Reza Kourdy
Department of Computer Engineering Islamic Azad University KhorramAbad Branch, Iran

Abstract this paper presents some experiences with Class of Service (CoS) translation in networks-on-Chip. Multi Protocol Label Switching(MPLS) provides the flexibility of being able to Forwarding Equivalence Classes and the ability to create a forwarding hierarchy via label stacking. All of these techniques facilitate the operation of Quality of Service. Through MPLS-TE (Traffics Engineering) functions such as network resources optimization, strict Quality of Service voice data delivery, and fast recovery on link or node failures can ensured. We simulate Butterfly NoC architecture with Network Simulator 2 (NS2). The simulation results reveal the applicability of the proposed architecture, which can support classes of services with simplicity of internetwork connection of Butterfly architecture. Keywords- Network-on-Chip, MultiProtocol Label Switching, Quality-of-Service , claccess of services.

integrate the traditional Internet Protocol (IP) routing and switching technologies because it provides end-to-end Quality of Service (QoS), guarantees Traffic Engineering, and support Virtual Private Networks (VPNs). However, MPLS must use path restoration schemes to guarantee the delivery of packets through a network.[6] Multi-Protocol Label Switching (MPLS) is an alternative to integrate the traditional Internet Protocol (IP) routing and switching technologies because it provides end-to-end Quality of Service (QoS), guarantees Traffic Engineering, and support Virtual Private Networks (VPNs). However, MPLS must use path restoration schemes to guarantee the delivery of packets through a network [7]. The MPLS Traffic Engineering Router Architecture shown in Fig. 1.

INTRODUCTION

Systems-on-chip (SoCs) for multimedia or telecommunication applications will contain a large number of processing elements (PEs) such as a DSP processor, RISC CPU, embedded RAM, graphics engine, etc. As a result, there is a need for high-throughput communications links between these blocks. There exist many bus based SoCs which are widely used in industry such as AMBA [1], IBM Core-Connect [2], Pi-Bus [3], etc. During the last years, Networks-on-Chip (NoCs) have become a true alternative for the design of complex integrated Systems-on-Chip (SoC). Much effort has been spent for research on functionalities, mechanisms, and Quality-of-Service (QoS) features in NoCs [4]. We can view an NoC system platform as a very complex and multiprotocol network. As such, providing application-wide, endto-end quality-of-service (QoS) is crucial for optimum system performance. System-wide performance constraints require predictability of inter-block communication and QoS guarantees for the end-to-end communication. QoS is characterized by diverse parameters, such as reliability, delay, jitter, bandwidth, packet loss, and throughput [5]. For this purpose, there are recovery mechanisms for MPLS-based protection label switching path (LSP), which are protection switching and rerouting models. MultiProtocol Label Switching (MPLS) is an alternative to

Fig. 1. MPLS-TE Router Architecture.

BACKGROUND

2.1 Bandwidth Reservation System designers will require NoC communication platforms which have a certain degree of multi-protocol support such as MPLS. Most of the existing NoC architectures are packet-switched (connectionless) NoCs.

JOURNAL OF COMPUTING, VOLUME 4, ISSUE 4, APRIL 2012, ISSN 2151-9617 https://sites.google.com/site/journalofcomputing WWW.JOURNALOFCOMPUTING.ORG

96

They are targeting Best-Effort (BE) traffic. Architectures offering only Best-Effort services dont reserve bandwidth and hence can have better average resource utilization, at the cost of unpredictable or unbounded worst-case behavior [8]. 2.2 Using MPLS For Guarantee QOS in NOCs The Multi Protocol Label Switching (MPLS) technology has emerged as a connection-oriented protocol serving connectionless internet IP networks, and thus it provides the means for traffic engineering. This means that paths are set up for aggregated flows of a certain type between specified end points of the IP traffic. Also, DiffServ can support a scalable QoS. The combination of these two approaches leads to a scalable hard QoS on internet IP networks because MPLS creates paths that can be traffic engineered [9]. The MPLS Node Architecture for supporting Quality of services has been shown in Fig. 2.

In order to support MPLS real-time traffic, the Service Classifier component has been designed and implemented CBQ (Class Based Queuing), which was implemented on NS, is selected for the Packet Scheduler component.

RELATED WORK

Network-on-Chip Quality-of-Service through MPLS objective is discussed in [10], [11] and just papers that use MPLS in NOCs.

SYSTEM ARCHITECTURE

Fig. 2. The multilayer integrated QoS control architectural model in MPLS-NoCs.

Another scheme for process the traffic in order to supporting quality of services with Differentiated Services (DiffServ) shown in fig 3.

4.1 MPLS Components and terms This section gives an overview of the terms associated with the MPLS technology. 1) Path Label (Associates a packet to a FEC) FEC(Forward Equivalence Class) : A subset of packets that are all treated the same way by an LSR LSP (Label Switch Path): path that a packet follows for a specific FEC Shim - Header containing a Label Stack Label Stack - Multiple labels containing information on how a packet is forwarded. 2) Routers LSR(Label Switched Router)An MPLS capable router LER(Label Edge Router) 3) Protocol LDP (Label Distribution Protocol), used to distribute Label information between MPLS-aware network devices CR-LDP (constraint-based LDP) contains extensions for LDP to extend its capabilities. This allows extending the information used to setup paths beyond what is available for the routing protocol 4) MNS(MPLS Network Simulator) Mpls-Node PDU(Protocol Data Unit) 4.2 MPLS-NoC Architecture MPLS-NoC Architectures are composed of following components: 5) Label-Edge Router The Label Edge Router (LER) operates at the edge of an MPLS network. It is necessary to include an interface (IF) module to interact with different semi-global SoC/NoC [12]. 6) Label-Switched Router A Label Switch Router (LSR) is the main component of the MPLS network. It sets up a path to other MPLS routers and forwards packets to them. 7) Label-Switched Path(LSP) Within an MPLS domain, a path is set up for a given packet to travel on an FEC. The Label-Switched Path (LSP) is set up prior to data transmission. 8) MPLS-NOC MODELING

Fig. 3. Difserv architecture in MPLS NOCs.

JOURNAL OF COMPUTING, VOLUME 4, ISSUE 4, APRIL 2012, ISSN 2151-9617 https://sites.google.com/site/journalofcomputing WWW.JOURNALOFCOMPUTING.ORG

97

4.3 MPLS Operation The function of ingress LER is to put a label in the IP packet and forward it to the next hop in the MPLS network. This label is assigned according to the forwarding equivalence class (FEC) of the packet. In this case the IP packet is encapsulated in an MPLS PDU, with an MPLS shim header included in the packet. The main objectives of MPLS are accomplished using fixed-length labels. These labels included in an MPLS header are assigned considering FECs that determine the route of a datagram. The FECs are a representation of a group of packets that share the same requirements to their transport. These FECs can be used to support QoS operations (e.g. real time applications) [13]. This FEC to label relationship determine the Label Switched Path (LSP) of a datagram, from the ingress point to the egress point of the MPLS network. The LSP or tunnel at both ends of the MPLS network is a concatenation of the LSP segments between each node. In this tunnel the ingress node define the type of traffic and assigns a label. According to this label, the traffic is forwarded through the LSP without further examination. At the end of the tunnel, the egress node removes the label and forwards the traffic to an external network (e.g. an IP network). This type of tunnels allows the implementation of Traffic Engineering (TE) [6].

Fig. 4. ButterFly NoC architecture with IP packet forwarding.

The square was processor elements and the circle and hexagon elements were switches that connect to each other. Our communication was MPLS based with types of traffics (see fig. 5).

EVALUATIONS

5.1 Simulation Framework In this paper, we have modeled our MPLS-NoC architecture concepts with the widely used network simulator ns-2 [14]. NS-2 has been widely applied in research related to the design and evaluation of computer networks and to evaluate various design options for NoC architectures [15], including the design of routers, communication protocols, etc. 5.2 Simulation Results We use MPLS types of quality of services in Network on chip in Butterfly topology that shown in Fig. 4 and Fig. 5 respectively. Four common classes of services that we can supports by MPLS mechanism for NOCs are: RTS as Real time services STS as Streaming services HBES as High Best Effort services BES as Best effort services In IP NOCs, just we can consider Best Effort services for traffic but in Mpls-NOCs, we have more services that mentioned previously. Thus we can support types of Quality of Services for our communication in Network on chip.

Fig. 5. ButterFly NoC with MPLS classes of services .

As shown in fig. 5, our communication was between node 7 and node 10 that was that was connected to switch 0 and 3 respectively. The communication path was one path, with the name 0_3 label switching path (LSP). We consider that the network link was homogeneously, and the links between switches has one Megabit/sec. The bandwidth for Real traffic (RTS) is set to be 700 Kilobit/sec; Streaming services (STS) generate the constant bit rate (CBR) traffic with 250 Kilobit/sec which starts at time 10, High Best Effort services (HBES) generate the CBR traffic with 600 Kilobit/sec, Best effort services (BES) generate the constant bit rate (CBR) traffic with 250 Kilobit/sec, that showed in Fig. 6. Total rate of all traffics in 0_3 LSP before of the time 10 is equals to (250 kbps (STS) + 600 (HBES) kbps + 250 (BES) kbps = 1.1 Mbps). Thus the communication load between

JOURNAL OF COMPUTING, VOLUME 4, ISSUE 4, APRIL 2012, ISSN 2151-9617 https://sites.google.com/site/journalofcomputing WWW.JOURNALOFCOMPUTING.ORG

98

mentioned cores is more than the bandwidth in the path 0_3 LSP (fig. 6), and we have 100 kbps overflow traffics. Because of lack of priority between types of traffics, In IP based NoCs this overflow traffics was dropped from all traffics, but in MPLS- NoCs we have priority between all mentioned types of traffics. The lowest priority was related to Best Effort Services (BES). Thus this type of traffics caused to drop and remaining bandwidth for Best Effort Services (BES) was equals to (250 kbps-100 kbps=150 kbps).

platform that supports several classes of services to support QoS with multi-protocol concepts, which were adapted from the Multi Protocol-Level Switching technique with DiffServ.

7
[1] [2] [3]

REFERENCES
J. Park, I. Kim, S. Kim, S. Park, B. Koo, K. Shin, K. Seo, and J. Cha, MPEG-4 video codec on an ARM core and AMBA, in Proc. Of Workshop and Exhibition on MPEG-4, Jun. 2001, pp. 9598. R. Hofmann and B. Drerup, Next generation CoreConnect processor local bus architecture, in 15th Annual IEEE International ASIC/SOC Conference, Sep. 2002, pp. 221225. C. Roark and F. Jackson, New developments in a PI-Bus specification by the JIAWG and SAE, in Proc. of the IEEE National Aerospace and Electronics Conference, vol. 2, May 1992, pp. 760 766. S. Kubisch, C. Cornelius, R. Hecht, D. Timmermann, Mapping a Pipelined Data Path onto a Network-on-Chip, University of Rostock, Institute of Applied Microelectronics and Computer Engineering, 4-6 July 2007 Industrial Embedded Systems, 2007. M. Harmanci, N. Escudero, Y. Leblebici, and P. Ienne, Quantitative modelling and comparison of communication schemes to guarantee quality-of-service in networks-on-chip, in IEEE International Symposium on Circuits and Systems, vol. 2, May 2005, pp. 1782 1785. M. Minero-Muoz, V. Alarcon-Aquino, Reconfigurable Path Restoration Schemes for MPLS Networks, Infocomp Journal of Computer Science, Vol. 8, No2, June 2009; pp. 29-38. M. MINERO-MUOZ, V. ALARCON-AQUINO, Reconfigurable Path Restoration Schemes for MPLS Networks ,Department of Computing, Electronics, and Mechatronics Communications and Signal Processing Research Group UDLAP - Sta. Catarina Martir, Cholula Puebla MEXICO, April 26, 2009. K. Goossens, J. Dielissen, J. van Meerbergen, P. Poplavko, A. Rdulescu, E. Rijpkema, E. Waterlander, and P. Wielage, Guaranteeing the quality of services in networks on chip, pp. 61 82, 2003. V. Fineberg, C. Chen, and X. Xiao, An end-to-end QoS architecture with the MPLS-based core, in IEEE Workshop on IP Operations and Management, 2002, pp. 2630. M. Kim, D. Kim and G. E. Sobelman, Network-on-Chip Quality-ofService through MultiProtocol Label Switching Department of Electrical and Computer Engineering, University of Minnesota, Minneapolis, MN 55455 USA. M.R.N, R.Kourdy, M.R.Nouri Rad, "Network-on-Chip Quality-ofService through MPLS Reservation Mechanism ", 2010 Second International Conference on Computer Research and Development, Kuala Lumpur, May 2010, pp. 775 778. R. Peterkin and D. Ionescu, Embedded MPLS Architecture, in Proc. of the 19th IEEE International Parallel and Distributed Processing Symposium, Apr. 2005, pp. 170a170a. Black, U. MPLS and Label Switching Networks. Prentice Hall PTR, Upper Saddle River, NJ, USA, 2002. www.isi.edu/nsnam/ns R. Lemaire, F. Clermidy, Y. Durand, D. Lattard, and A. Jerraya, Performance Evaluation of a NoC-Based Design for MC-CDMA Telecommunications Using NS-2, in The 16th IEEE International Workshop on Rapid System Prototyping, Jun. 2005, pp. 2430.

[4]

[5]

[6] [7]

Fig. 6. Bandwidth utilization with classes of services.

As shown in fig. 6, at time 10 the real traffic (RTS) started, but the traffic that this types of traffic needs, with addition other traffics is equals to (700 kbps (RTS) +250 kbps (STS) + 600 (HBES) kbps + 150 (remaining BES) kbps = 1.7 Mbps). Thus we have 700 kbps more traffics in communication path. Real traffic (RTS) gets its remaining bandwidth from Best Effort (BES) and High Best Effort (HBES) traffic that have less priority than real traffic (RTS). The High Best Effort traffic (HBES) has more priority than Best effort traffic (BES). As a result we can conclude that: a) 700 kbps -150 (remaining BES) kbps=550 kbps remains for Real Traffic (RTS), and communication load for Best Effort traffic (BES) reaches to zero. b) 600 (HBES) kbps - 550 kbps (RTS needs) =50 kbps remains for High Best Effort traffic (HBES). At time 30 the real traffic (RTS) stopped and the other traffics return to previous status.

[8]

[9] [10]

[11]

[12] [13] [14] [15]

CONCLUSION AND FUTUREWORK

This paper presents some of the results obtained by using MPLS-NoC CoS with relative and fixed bandwidth allocation to Multi Protocol Label Switching Network on Cip. We have to present a novel NoC communication

Você também pode gostar