KR20090053180A - Queuing method by virtual clock, and scheduling apparatus and method using that - Google Patents

Abstract

The present invention relates to a queuing method, a scheduling apparatus using the same, and a computer readable recording medium having recorded thereon a program for realizing the method, and the packet arrival rate (especially, the average arrival rate of packets) for non-real time traffic. Queuing using virtual time, a scheduling apparatus using the same, and a method thereof, and each of the above methods for queuing and scheduling by using a virtual time and a virtual time to minimize packet discard and to guarantee quality of service (QoS) of non-real-time traffic. To provide a computer readable recording medium having recorded thereon a program for realizing this.

To this end, the present invention provides a scheduling apparatus comprising: packet classification means for classifying non-real-time traffic according to a packet arrival rate; Virtual time applying means for applying a virtual time according to the packet arrival rate for each traffic classified by the packet classification means; Queuing means for adjusting a queuing policy based on a virtual time for the packet to which the virtual time is applied in the virtual time applying means; And virtual time scheduling means for scheduling the packet queued by the queuing means based on the virtual time.

Queuing, scheduling, non-real-time traffic, packet arrival rate (packet average arrival rate), virtual time, secondary virtual time, critical interval

Description

Queuing method using virtual time, scheduling apparatus using same and method therefor {Queuing method by virtual clock, and scheduling apparatus and method using that}

The present invention implements a queuing method, a scheduling apparatus using the same, a method thereof, and each method for classifying non-real-time traffic inputted from a terminal or the like in a synchronous Ethernet, etc., and for queuing and scheduling the classified non-real-time traffic. The present invention relates to a computer-readable recording medium that records a program to be programmed. More specifically, the synchronous Ethernet classifies traffic inputted from a terminal during a given transmission cycle, and secures bandwidth through a known reservation method for real-time traffic. And queuing and scheduling the packet arrival rate (particularly, the average arrival rate of the packet) and the virtual time for the non-real-time traffic in the remaining bandwidth, thereby minimizing packet discard and guaranteeing QoS of the non-real-time traffic. Queuing method Using the present invention relates to a computer-readable recording medium recording a program for realizing a scheduling apparatus and method and the respective methods.

In the following embodiment, the present invention is described as an example applied to synchronous Ethernet (Residential Ethernet, Synchronous Ethernet, Audio / Video Bridging) dealing with real-time traffic and non-real-time traffic in an Ethernet environment as an example, the present invention is Note that it is not limited.

Ethernet technology is the most widely installed technology with over 80% market share of LAN technology. Ethernet technology was originally developed by "Xerox" and was developed by "Xerox", "DEC" and "Intel" and is now defined as a standard in Institute of Electrical and Electronics Engineers (IEEE) 802.3.

Conventional Ethernet technology has competitively accessed shared media using the Carrier Sense Multiple Access / Collision Detect (CSMA / CD) protocol as defined in IEEE 802.3. In other words, the Ethernet technology follows the CSMA / CD scheme, which gives the same priority to all traffic and transmits it through the competition, which is not suitable for video or voice transmission that is sensitive to transmission time delay.

Accordingly, a technique of IEEE 802.1Q has been proposed as a method for reducing delay for data having high priority such as multimedia data in Ethernet. IEEE 802.1 Q technology improves the transmission delay to some extent by first transmitting higher priority data than the existing IEEE 802.3 Ethernet technology.

However, since the IEEE 802.1Q technology operates in a non-preemptive manner, transmission may be delayed due to low priority data, and the transmission delay increases if the transmission distance is far.

Accordingly, synchronous Ethernet technology has emerged to solve the above problems while using the existing Ethernet. This synchronous Ethernet technology is currently used to synchronize devices among IEEE 802.1 Audio Video Bridging Task Groups (AVB TGs), to reserve bandwidth for real-time traffic, and to properly transmit real-time and non-real-time traffic. There are three major parts of the technology, and standardization is underway.

Among the technologies for transmitting real-time and non-real-time traffic, which are currently being standardized by IEEE 802.1 Qav, in case of real-time traffic, the reserved bandwidth is used to transmit the traffic, and for non-real-time traffic, the traffic is managed by various queuing scheduling. Transmitting. Here, SRP (Stream Reservation Protocol) is currently considered as a technology for reserving bandwidth for real-time traffic. In addition, WFQ (Weighted Fair Queuing), which is a typical queuing scheduling method, is used to transmit non-real-time traffic.

1 is a view for explaining a weighted fair queuing (WFQ) operation according to the prior art.

The WFQ technology was developed to solve the problem of low priority packets, which is a disadvantage of priority queuing, being killed by high priority traffic, while also failing to provide differentiated services in custom queuing. to be.

At this time, since the WFQ technique that can use a maximum of 4096 (N = 4096) queues 13 provides a large number of queues, the classes are classified according to the characteristics of each packet, and the classified classes are flow units 11 Use divided by. The WFQ scheduling scheme avoids queue congestion through a tail drop 12 which discards all packets arriving after the queue 13 is full.

Each queue 13 of the WFQ configured in the FIFO (First In First Out) method is weighted based on an Internet Protocol (IP) priority. More precisely, they are weighted inversely proportional to the size of the IP priority. The scheduling is performed to serve based on the actual packet size and the size of the virtual packet calculated using a known virtual packet calculation formula.

The reason for applying such a known virtual packet calculation formula is to preferentially service packets of high priority or small size. In other words, after classifying all traffics with different bandwidth requirements by flows, we tried to guarantee QoS of traffics through priority and weight assignment.

However, even in the WFQ scheduler 14, if the continuous delay continues and the queue fills up, traffic may be congested. Accordingly, there is a problem in that a packet is discarded if a high-priority traffic having a high bandwidth requirement continues to accumulate a service delay even with a reallocated bandwidth according to a weight.

Accordingly, there is a need for a research on an efficient queuing scheduling method capable of minimizing packet discard to guarantee QoS.

2 is an explanatory diagram for a queuing scheduling method for guaranteeing a quality of service (QoS) according to the prior art.

As shown in FIG. 2, the classifier 22 according to the related art classifies (classifies) the input packet 21 in units of flows based on a classification criterion with reference to the header information of the input packet 21. At the same time, determine the class of traffic to which the flow belongs. Here, determining the class of traffic means that the input traffic is actually queued and how it is scheduled.

At this time, the packet header information referred to by the classifier 22 includes a source IP address, a destination (destination) IP address, a protocol ID field, transmission port numbers of transmission control protocol (TCP) and user datagram protocol (UDP), There are destination (destination) port numbers for TCP and UDP. These are called 5 tuples and are fields that can distinguish IP flows.

In the IEEE 802.1Q frame structure of synchronous Ethernet, on the other hand, the input port number field and the output port number field are classification fields that distinguish one layer, and a destination (destination) address (DA) field and a source address (SA). Field, the EtherType (EtherType) field, the VLAN (virtual LAN) ID field, and the IEEE 802.1p / Q field are classification fields that distinguish two layers. In addition to the tuples specified in the classifier 22, the third layer uses TOS (Type Of Service), IP Priority (Precedence), and DSCP (DiffServ Code Point) as classification fields. Finally, in the fourth layer, a flag field for distinguishing TCP and UDP, a source port number field, and a destination port number field may be classified as classification fields.

Each packet 23 passing through the classifier 22 is characterized by a meter 24 assigned to each traffic flow. The measured results are compared with the QoS traffic characteristics promised in advance, and the markers 25 are marked at several priorities according to the comparison result, and polishing and shaping according to a predetermined policy. (Shaping) (26). The packet then exits the queue via the queuing scheduler 28 after passing through a queue manager 27 which deals with how the queue is handled upon arrival of a packet that exceeds its internal processing capacity.

However, the conventional method as described above cannot satisfy the QoS between end users due to scheduling delay when processing the packet based on packet priority or DSCP (DiffServ Code Point) to handle time delay sensitive traffic. There is a problem that a packet is discarded due to a scheduling delay.

That is, the prior art as described above has a problem in that QoS between end users cannot be satisfied due to a scheduling delay, and a packet is discarded due to a cumulative scheduling delay. It is an object of the present invention to solve such a problem.

Accordingly, the present invention queues and schedules packet arrival rate (especially, average arrival rate of packets) and virtual time for non-real-time traffic, thereby minimizing packet discard and ensuring quality of service (QoS) of non-real-time traffic. It is an object of the present invention to provide a queuing method using a virtual time, a scheduling apparatus using the same, and a computer readable recording medium having recorded thereon a program for realizing the method.

That is, the present invention divides non-real-time traffic using packet arrival rate (especially, average arrival rate of packets) in synchronous Ethernet, which handles real-time traffic and non-real-time traffic in Ethernet environment, and queues and schedules using virtual time. Queuing method using virtual time, a scheduling device using the same, and a program for realizing the above methods and a program for realizing each method can be read by a computer for minimizing packet discard and ensuring quality of service (QoS) of non-real-time traffic. The purpose is to provide a recording medium.

The objects of the present invention are not limited to the above-mentioned objects, and other objects and advantages of the present invention which are not mentioned above can be understood by the following description, and will be more clearly understood by the embodiments of the present invention. Also, it will be readily appreciated that the objects and advantages of the present invention may be realized by the means and combinations thereof indicated in the claims.

An apparatus of the present invention for achieving the above object comprises: a packet classification means for classifying non-real-time traffic according to packet arrival rate; Virtual time applying means for applying a virtual time according to the packet arrival rate for each traffic classified by the packet classification means; Queuing means for adjusting a queuing policy based on a virtual time for the packet to which the virtual time is applied in the virtual time applying means; And virtual time scheduling means for scheduling the packet queued by the queuing means based on the virtual time.

On the other hand, the method of the present invention for achieving the above object, the scheduling method, packet classification step of classifying the non-real-time traffic according to the packet arrival rate; A virtual time applying step of applying a virtual time according to a packet arrival rate for each classified traffic; A queuing step of adjusting a queuing policy for the packet to which the virtual time is applied based on the virtual time; And a virtual time scheduling step of scheduling the queued packet based on the virtual time according to the queuing policy.

On the other hand, another method of the present invention for achieving the above object, in the queuing method, by comparing the current virtual time and the secondary virtual time of the incoming packet by applying the current virtual time (timestamp) to the current time and the risk interval Adjusting a queuing policy; And discarding the packet with a tail drop or inserting the packet according to the adjusted queuing policy.

On the other hand, the present invention, a scheduling apparatus having a processor, packet classification function for classifying non-real-time traffic according to the packet arrival rate; A virtual time application function for applying a virtual time according to a packet arrival rate for each classified traffic; A queuing function of adjusting a queuing policy for the packet to which the virtual time is applied based on the virtual time; And a computer-readable recording medium having recorded thereon a program for realizing a virtual time scheduling function for scheduling a queued packet based on a virtual time according to the queuing policy.

In addition, the present invention, the function to adjust the queuing policy by comparing the current virtual time and the auxiliary virtual time of the incoming packet with the current virtual time (time stamp) is applied to the queuing device having a processor ; And a computer readable recording medium having recorded thereon a program for realizing a function of inserting a packet or discarding the packet by tail drop according to the adjusted queuing policy.

As described above, the present invention queues and schedules packet arrival rate (particularly, average arrival rate of packets) and virtual time for non-real time traffic, thereby minimizing packet discard and improving quality of service (QoS) of non-real time traffic. There is a guaranteeable effect.

That is, the present invention divides non-real-time traffic using packet arrival rate (especially, average arrival rate of packets) in synchronous Ethernet, which handles real-time traffic and non-real-time traffic in Ethernet environment, and queues and schedules using virtual time. There is an effect of minimizing packet discard and guaranteeing the quality of service (QoS) of non-real-time traffic.

In addition, the present invention classifies traffic inputted from a terminal during a given transmission cycle in synchronous Ethernet, etc., and secures bandwidth through a known reservation scheme in the case of real-time traffic, and packet arrival rate and virtual time for non-real-time traffic in the remaining bandwidth. In the case of queuing and scheduling using, since the order in which packets are serviced can be determined according to the packet arrival rate, each flow can be fairly processed in the intermediate bridge, thereby minimizing packet discard. .

In addition, the present invention has an effect that can be controlled through the monitoring process when burst traffic is applied to the intermediate bridge.

The above objects, features, and advantages will become more apparent from the detailed description given hereinafter with reference to the accompanying drawings, and accordingly, those skilled in the art to which the present invention pertains may share the technical idea of the present invention. It will be easy to implement. In addition, in describing the present invention, when it is determined that the detailed description of the known technology related to the present invention may unnecessarily obscure the subject matter of the present invention, the detailed description thereof will be omitted. Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

3 is a diagram illustrating a queuing method using virtual time in synchronous Ethernet, a scheduling apparatus using the same, and a method thereof, and an IEEE 802.1 Q frame structure in synchronous Ethernet according to an embodiment of the present invention.

As shown in FIG. 3, the first packet classifier 302 that classifies non-real-time traffic in synchronous Ethernet uses four tuples, unlike a conventional classifier (22 in FIG. 2) that uses five tuples. Frames used to transmit data and messages in synchronous Ethernet use IEEE 802.1 Q frames because they cover two layers. In the first packet classifier 302 of the synchronous Ethernet, the traffic 301 that enters (inputs) from a terminal or the like is transmitted to a source address, a destination (destination) address, a VLAN (Priority LAN) priority, and an ether type ( Classify non-real-time traffic using EtherType, Ethernet type). At this time, the traffic divided into real-time traffic is processed in a traffic processing method through the reservation, which is a known method.

As shown in FIG. 3, the IEEE 802.1 Q frame used for data and message transmission in an IEEE 802.1 AVB LAN includes a destination (destination) address (DA) field 308, a destination address (SA) field ( 309), Tag Protocol IDentifier (TPID) field 310, Virtual LAN (VLAN) ID field 311, EtherType (EtherType, Ethernet) field 312, Data field 313, and Frame Check Sum) field 314.

The destination (destination) address (DA) field 308 stores a destination address to receive the frame, and the sender address (SA) field 309 stores a destination address that transmitted the frame. Immediately after the source address (SA) field 309 is a TPID field 310, which serves to indicate the presence of a VLAN tag.

There are three detailed fields in the VLAN ID field 311, a priority field 315 indicating one of eight priorities of IEEE 802.1Q, a CFI indicating the format of a transmission medium, that is, Ethernet or token ring. (Canonical Format Identifier) field 316, and a VID field 317 for recording the actual VLAN ID for distinguishing a VLAN.

In addition, the IEEE 802.1 Q frame includes an EtherType (EtherType) field 312 indicating a protocol of a higher layer and a data field 313 into which actual data is inserted, and the FCS field 314 corresponds to a corresponding frame. Contains a special 4-byte code for error checking.

Among these fields, the present invention enters the bridge using the destination address field 308, the source address field 309, the VLAN priority field 315, and the EtherType (EtherType) field 312. Traffic 301 is identified. That is, the first packet classifier 302 classifies (classifies) non-real-time traffic by EtherType (type of Ethernet) and groups the traffic by VLAN priority (303).

The second packet classifier 304 secondarily classifies (classifies) non-real-time traffic classified primarily by the first packet classifier 302 according to the packet arrival rate. In this case, it is preferable to use an average arrival rate (AR) of the packet as the arrival rate. In addition, the packet arrival rate may be used, or the packet arrival rate may be implemented by using the packet size or priority used in the conventional scheme.

In the following exemplary embodiment, a case in which an average arrival rate (AR) of a packet is used as the packet arrival rate will be described as an example. However, the present invention is not limited thereto. As such, when the average arrival rate (AR) of the packet is used as the packet arrival rate, the second packet classifier 304 may be implemented as an average arrival rate (AR) classifier.

The marker 305 marks the timestamp using the virtual time for each traffic classified by the second packet classifier 304. At this time, an AR meter included in the marker 305 calculates a timestamp using virtual time for each traffic classified by the second packet classifier 304. Here, the timestamp (current virtual time) is a value obtained by adding Vtick (Vtick = 1 / average arrival rate) to the previous virtual time.

In addition, the queue manager 306 allocates each packet marked with a time stamp (current virtual time) to the corresponding queue based on the virtual time in the marker 305, and when the corresponding queue becomes full during such queuing, Perform a tail drop to discard the packet.

The virtual time scheduler 307 extracts and transmits a packet from the queue based on the virtual time during normal queuing.

FIG. 4 is a diagram illustrating an embodiment of a scheduling apparatus using virtual time in synchronous Ethernet according to the present invention, and shows an overall configuration of an apparatus for scheduling non-real-time traffic using virtual time in synchronous Ethernet.

As shown in FIG. 4, a scheduling apparatus using virtual time in synchronous Ethernet according to the present invention includes a first packet classifier 402 and a first packet classifier 402 for classifying (dividing) non-real-time traffic. A second packet classifier 404 and a second packet classifier 404 for classifying the non-real-time traffic classified primarily according to packet arrival rate (eg, average arrival rate of packets). A queuing policy based on the virtual time for the packet introduced by the AR meter 405 and the marker 406 for marking the timestamp using the virtual time, and the timestamp (current virtual time) marked by the marker 406 for each time. Scheduling the packet based on the virtual time of the queue manager 409 for managing the queues 411 to 414, and the packets queued to the queues 411 to 414 by the queue manager 409. For virtual A time scheduler 410.

Next, the above components will be described in more detail.

First, packets 401 inputted from a terminal or the like are first classified (classified) and grouped into non-real-time traffic by the first packet classifier 402 (403), and the non-real-time traffic thus classified is classified into a second packet classifier ( AR classifier (404) is again divided by the average arrival rate of the packet. The flows divided by the average arrival rate of the packets are marked with a time stamp 407 corresponding to each flow by a meter defined as an AR meter 405 and a marker 406. That is, the AR meter 405 and the marker 406 have a time stamp (current virtual arrival rate) to which Vtick (1 / average arrival rate) is added based on the virtual time 408 for each of the flows divided by the average arrival rate of the packets. Time (407) is allocated and marked. That is, the current virtual time (timestamp) of the newly generated packet is calculated as shown in Equation 1 below.

Current virtual time (timestamp) = previous virtual time + Vtick

Where Vtick is " 1 / average arrival rate ".

At this time, flows with a high average arrival rate have a small Vtick and flows with a small average arrival rate have a large Vtick, so that a flow having a high average arrival rate has a low queuing delay.

The queue manager 409 adjusts a queuing policy by comparing a current virtual time and a current time of a packet introduced by marking a time stamp (current virtual time) at the marker 406. That is, the queue manager 409 adjusts a queuing policy by comparing a current virtual time and an auxiliary virtual time of a packet introduced by marking a timestamp (current virtual time) at the marker 406 with a current time and a critical section. The packet is inserted into the queues 411 to 414 according to the adjusted queuing policy. The queue manager 409 discards the packet with a tail drop policy to prevent an overload of the system when an overflow possibility condition in the queues 411 to 414 occurs.

The virtual time scheduler 410 schedules and transmits the packet based on the virtual time of the packet queued in the queues 411 to 414. In FIG. 4, packets 4, 5, and 6 of the highest queue 411 are first extracted and transmitted, and then packets 10 and 10 of the second queue 412 and 10, of the third queue 413 After extracting and transmitting, packets 12 and 14 of the second queue are extracted and transmitted again, and finally, scheduling is performed in the order that packets 30 of the lowest queue 414 are extracted and transmitted.

5A to 5C are diagrams for describing a queuing policy using virtual time in synchronous Ethernet according to the present invention. FIG. 5A to 5C are queuing markers and queue managers managing non-real-time traffic for each classified flow in synchronous Ethernet. It is a figure for demonstrating a policy.

The queues shown in FIGS. 5A-5C are virtual queues (created by priorities) and represent queues shown in the marker 406 of FIG. 4.

Referring to FIG. 5A, in the virtual queue 501 of FIG. 5A, when the current virtual time 503 is earlier than the current time 502 and less than the danger section 504, the queue manager 409 may be normally queued. Do it. At this time, the current time is the time that the packets entering the queue are processed, the critical section 504 at the point of 90% of each queue is a reference point to prevent the system from being overloaded due to the overflow of the queue is full. , The current virtual time is a value calculated according to Equation 1 above.

Referring to FIG. 5B, in the queue 505 of FIG. 5B, the queue manager 409 becomes a tail drop policy from the moment when the current virtual time 507 is greater than the current time 506 but equals the danger section 508. Discard the packet so that it no longer enters the queue.

Referring to FIG. 5C, the queue 509 of FIG. 5C is a case where the current virtual time 510 is smaller than the current time 511. This indicates a case where a packet keeps coming in and does not come in during an Average Observation Interval Range (AIR) or when a packet comes in later than the speed at which the current time passes. Here, the average observation interval (AIR) represents the total amount of packets entering the product of the average arrival rate (AR) of the packet and the average observation interval (AI) in which the packet is continuously serviced to the average observation interval of the packet. The marker 406 compares the present time with the present virtual time for each first section of the mean observation section AIR. The marker 406 uses the auxiliary virtual time 512 when the current virtual time 510 is less than the current time 511, such as the queue 509 of FIG. 5C. Accordingly, the virtual time scheduling operation can be continued by adding the newly measured Vtick by the average arrival rate of the packet entering the secondary virtual time 512 updated to the current time even if the packet enters again after the average observation interval AIR. Will be. The virtual time scheduling task allows to continue queuing and scheduling packets until the current virtual time 510 is not greater than or equal to the critical interval 513.

6 is a flowchart illustrating a queuing method using virtual time in synchronous Ethernet and a scheduling method using the same according to the present invention, and a queuing method using virtual time for guaranteeing QoS of non-real-time traffic in synchronous Ethernet and using the same A flowchart for explaining a scheduling method.

First, when a packet arrives for the first time, a marker (including an AR meter) sets the virtual time for each flow and the auxiliary virtual time to the current time. As soon as each packet arrives in the virtual queue, a marker (including an AR meter) sets the packet's current virtual time (timestamp) by adding Vtick (1 / average arrival rate) to the previous virtual time. Likewise, markers (including the AR meter) add Vtick to the secondary virtual time. The secondary virtual time then has the greater of the current time and the secondary virtual time. Here, observe the average arrival rate (AR) of the packet and the average observation interval (AI) value in which the packet is continuously serviced by the average observation interval of the packet, and observe the fluid ingress of the packet as if the packets were bursted or not. The concept of virtual time and auxiliary virtual time is used to control it. Vtick is expressed as "average arrival rate of 1 / packet" as described above. In other words, flows with a high average arrival rate have a small Vtick value, which increases the frequency of queuing, and flows with a low average arrival rate have a large Vtick value, which results in a low queuing frequency.

As shown in FIG. 6, packets entering the bridge from an external device are first classified into two layers of traffic in the first packet classifier (601). In this case, the first packet classifier determines whether the input traffic is non-real-time traffic or real-time traffic by using an EtherType (EtherType, a type of Ethernet) (602).

As a result of the identification (division) 602, if the input traffic is real-time traffic, the corresponding traffic is processed by a known traffic processing method through reservation (614). If the input traffic is non-real-time traffic, the packets are grouped by priority. (603).

Thereafter, the second packet classifier classifies and classifies the non-real-time traffic classified by the group according to the packet arrival rate (604), and the packet classified by the group is allocated to an AR meter and a marker. As a result, corresponding timestamps are marked for each flow and inserted into the virtual queues (virtual queues are created for each priority) (605).

Thereafter, the queue manager compares the current virtual time (timestamp) with the current time (606). If the current virtual time is less than the current time, the queue manager queues using the auxiliary virtual time (607).

When the auxiliary virtual time is used as described above, it is checked whether the virtual time has been updated to the current time (608). If the virtual time has been updated to the current time, the process proceeds to step “610” and the queuing is performed again using the current virtual time. do. If the secondary virtual time is used and the virtual time is not updated to the current time, the secondary virtual time is compared with the dangerous section (609).

As a result of the comparison (609), if the auxiliary virtual time is greater than the critical period, the tail drop is applied to discard the packet to prevent queuing delay and proceeds to step "613" (612). On the contrary, if the auxiliary virtual time is smaller than the danger section, the process proceeds to “607” and continues queuing until the virtual time is updated to the current time using the auxiliary virtual time.

On the other hand, the comparison result 606, if the current virtual time is greater than the current time, the virtual time is compared with the danger interval (610), if the current virtual time is less than the danger interval, the queue continues by the priority and virtual time of each queue If the current virtual time is greater than the critical period (611), the queue manager discards the packet with a tail drop (612), and then proceeds to process "613".

The queue manager continuously queues and checks whether the virtual queue is empty (613). If the virtual queue is not empty, the queue manager returns to the moment when the packet enters the queue and compares the current virtual time with the current time again (606). If the virtual queue is empty, queuing and scheduling are terminated.

At this time, the packet is sent at the average arrival rate of packets entering the queue, but if the average arrival rate is the same, the packet is sent according to the priority of each sub-queue.

On the other hand, the method of the present invention as described above can be written in a computer program. And the code and code segments constituting the program can be easily inferred by a computer programmer in the art. In addition, the written program is stored in a computer-readable recording medium (information storage medium), and read and executed by a computer to implement the method of the present invention. The recording medium may include any type of computer readable recording medium.

The present invention described above is capable of various substitutions, modifications, and changes without departing from the technical spirit of the present invention for those skilled in the art to which the present invention pertains. It is not limited by the drawings.

The present invention may be used for synchronous Ethernet (Residential Ethernet, Synchronous Ethernet, Audio / Video Bridging), etc., which handles real-time traffic and non-real-time traffic in an Ethernet environment.

1 is a view for explaining a weighted fair queuing (WFQ) operation according to the prior art,

2 is a diagram illustrating a queuing scheduling method for guaranteeing a quality of service (QoS) according to the prior art;

3 is a diagram illustrating a queuing method using virtual time in synchronous Ethernet, a scheduling apparatus using the same, and a method thereof, and an IEEE 802.1 Q frame structure in synchronous Ethernet according to an embodiment of the present invention.

4 is a block diagram of an embodiment of a scheduling apparatus using virtual time in synchronous Ethernet according to the present invention;

5A through 5C are diagrams illustrating an exemplary queuing policy using virtual time in synchronous Ethernet according to the present invention;

6 is a flowchart illustrating a queuing method using virtual time and a scheduling method using the same in synchronous Ethernet according to the present invention.

* Explanation of symbols for the main parts of the drawings

402: first packet classifier 404: second packet classifier

405: AR Meter 406: Marker

409: queue manager 410: virtual time scheduler

Claims (26)

In the scheduling apparatus, Packet classification means for classifying non-real-time traffic according to packet arrival rate; Virtual time applying means for applying a virtual time according to the packet arrival rate for each traffic classified by the packet classification means; Queuing means for adjusting a queuing policy based on a virtual time for the packet to which the virtual time is applied in the virtual time applying means; And Virtual time scheduling means for scheduling a packet queued by the queuing means based on a virtual time Scheduling apparatus using a virtual time comprising a. The method of claim 1, The packet classification means, A first packet classifier for classifying non-real-time traffic; And A second packet classifier for classifying non-real-time traffic classified by the first packet classifier according to packet arrival rate Scheduling apparatus using a virtual time comprising a. The method of claim 2, The second packet classifier, A scheduling apparatus using virtual time, characterized in that the average arrival rate (AR) of the packet is used as the packet arrival rate. The method of claim 2, The second packet classifier, A scheduling apparatus using virtual time, characterized in that any one of a packet transmission rate, a packet size, and a priority is used as a packet arrival rate. The method according to any one of claims 2 to 4, The first packet classifier, A scheduling apparatus using virtual time, comprising classifying and grouping non-real-time traffic by using a source address, a destination (destination) address, a VLAN (Priority LAN) Priority, and an EtherType. The method of claim 5, wherein The first packet classifier, A scheduling apparatus using virtual time, characterized by classifying (classifying) non-real-time traffic by EtherType and grouping the traffic by VLAN priority. The method according to any one of claims 1 to 4, The virtual time applying means, Timestamp calculation means for calculating a timestamp (current virtual time) using the virtual time and the packet arrival rate for each traffic classified by the packet classification means; And Means for allocating and marking the timestamp (current virtual time) calculated by the timestamp calculation means for each classified traffic flow Scheduling apparatus using a virtual time comprising a. The method of claim 7, wherein The time stamp calculation means, The time stamp (current virtual time) is calculated by adding Vtick to the previous virtual time, wherein the Vtick is " 1 / packet average arrival rate ". The method of claim 7, wherein The queuing means, The queuing policy is adjusted by comparing the current virtual time and the auxiliary virtual time of the incoming packet with the timestamp (current virtual time) marked with the current time and the risk interval, and the packet is inserted or the tail is dropped according to the adjusted queuing policy. Scheduling apparatus using a virtual time, characterized in that for discarding packets. The method of claim 9, The queuing means, As the current virtual time (timestamp) is applied and the current virtual time of the incoming packet is larger than the current time and less than the critical section, queueing (packet insertion) is performed based on the virtual time. As the current virtual time is applied and the current virtual time of the incoming packet is greater than the current time and the critical section, the packet is dropped by tail drop. The scheduling apparatus using the virtual time, characterized in that for performing the queuing based on the auxiliary virtual time as the current virtual time of the packet introduced by the current virtual time is smaller than the current time. The method of claim 10, The queuing means, While queuing based on the auxiliary virtual time, it is checked whether the virtual time is updated to the current time, and when updated, the current virtual time is compared with a critical section to perform queuing (packet insertion) based on the virtual time or packet by tail drop. Discards the virtual time, and if it is not updated, compares the secondary virtual time with the critical section to continue queuing (packet insertion) based on the secondary virtual time or further discard the packet by tail drop. Scheduling device used. In the scheduling method, A packet classification step of classifying the non-real-time traffic according to the packet arrival rate; A virtual time applying step of applying a virtual time according to a packet arrival rate for each classified traffic; A queuing step of adjusting a queuing policy for the packet to which the virtual time is applied based on the virtual time; And Virtual time scheduling step of scheduling the queued packet based on the virtual time according to the queuing policy Scheduling method using a virtual time comprising a. The method of claim 12, The packet classification step, A first packet classification process for classifying non-real-time traffic; And A second packet classification process classifying the non-real-time traffic classified in the first packet classification process according to a packet arrival rate Scheduling method using a virtual time comprising a. The method of claim 13, The second packet classification process, A scheduling method using virtual time, characterized in that the average arrival rate (AR) of the packet is used as the packet arrival rate. The method of claim 13, The second packet classification process, A scheduling method using virtual time, wherein any one of a packet transmission rate, a packet size, and a priority is used as a packet arrival rate. The method according to any one of claims 13 to 15, The first packet classification process, A method for scheduling using virtual time, comprising classifying and grouping non-real-time traffic using a source address, a destination (destination) address, a VLAN (PLAN) Priority, and an EtherType. The method according to any one of claims 12 to 15, The virtual time applying step, A timestamp calculation step of calculating a timestamp (current virtual time) by using the virtual time and the packet arrival rate for each traffic classified in the packet classification step; And Assigning and marking the timestamp (current virtual time) calculated in the timestamp calculation process for each classified traffic flow Scheduling method using a virtual time comprising a. The method of claim 17, The timestamp calculation process, A timestamp (current virtual time) is calculated by adding Vtick to a previous virtual time, wherein the Vtick is " 1 / packet average arrival rate ". The method of claim 17, The queuing step, The queuing policy is adjusted by comparing the current virtual time and the auxiliary virtual time of the incoming packet with the timestamp (current virtual time) marked with the current time and the risk interval, and the packet is inserted or the tail is dropped according to the adjusted queuing policy. Scheduling method using a virtual time, characterized in that to discard the packet. The method of claim 19, The queuing step, Performing queuing (packet insertion) based on the virtual time as the current virtual time of the incoming packet is applied to the current virtual time (timestamp) than the current time and less than the dangerous interval; Discarding the packet with a tail drop as the current virtual time of the packet introduced by applying the current virtual time is greater than the current time and the danger interval; And Process of performing queuing based on the auxiliary virtual time as the current virtual time of the incoming packet is applied less than the current time by applying the current virtual time Scheduling method using a virtual time comprising a. The method of claim 20, The queuing method, Checking whether a virtual time is updated to a current time while queuing based on the auxiliary virtual time; As a result of the checking, when updated, comparing the current virtual time with a dangerous section to perform queuing (packet insertion) based on the virtual time or discarding the packet by tail drop; And As a result of the check, if the update is not updated, the secondary virtual time is compared with the critical section to continue queuing (packet insertion) based on the secondary virtual time, or discard the packet by tail drop. Scheduling method using a virtual time further comprising. In the queuing method, Adjusting a queuing policy by comparing a current virtual time and an auxiliary virtual time of an incoming packet by applying a current virtual time (timestamp) with a current time and a risk interval; And Inserting packets or discarding packets with tail drops according to the adjusted queuing policy Queuing method using a virtual time comprising a. The method of claim 22, The queuing method, Performing a queuing (packet insertion) based on the virtual time as the current virtual time of the incoming packet is applied by applying the current virtual time (timestamp) to be greater than the current time and less than the critical period; Discarding the packet with a tail drop as the current virtual time of the packet introduced by applying the current virtual time is greater than the current time and the danger interval; And Performing the queuing based on the auxiliary virtual time as the current virtual time of the incoming packet is applied than the current virtual time by applying the current virtual time Queuing method using a virtual time comprising a. The method of claim 23, The queuing method, Checking whether a virtual time is updated to a current time while queuing based on the auxiliary virtual time; As a result of the checking, when updated, comparing the current virtual time with a dangerous section to perform queuing (packet insertion) based on the virtual time or discarding the packet by tail drop; And If it is not confirmed, comparing the auxiliary virtual time with a critical section if not updated, continuously performing queuing (packet insertion) based on the auxiliary virtual time, or discarding the packet by tail drop; Queuing method using a virtual time further comprising. In a scheduling device having a processor, A packet classification function for classifying non-real-time traffic according to packet arrival rate; A virtual time application function for applying a virtual time according to a packet arrival rate for each classified traffic; A queuing function of adjusting a queuing policy for the packet to which the virtual time is applied based on the virtual time; And Virtual time scheduling function for scheduling a queued packet based on a virtual time according to the queuing policy A computer-readable recording medium having recorded thereon a program for realizing this. In a queuing device having a processor, A function of adjusting a queuing policy by comparing a current virtual time and an auxiliary virtual time of an incoming packet by applying a current virtual time (timestamp) with a current time and a risk interval; And A function of inserting a packet or discarding the packet by tail drop according to the adjusted queuing policy. A computer-readable recording medium having recorded thereon a program for realizing this.

Images