JP6061829B2 - Link quality measuring apparatus and flow entry aggregation method and program - Google Patents

Description

  The present invention relates to a link quality measuring apparatus, and a flow entry aggregation method and program thereof, and more particularly, to a link quality measuring apparatus capable of reducing the number of flow entries of a test packet registered in each switch on a network, and its flow entry aggregation method and Regarding the program.

  Non-Patent Document 1 discloses an architecture called SDN (Software Defined Network) for centrally controlling network devices with software. Non-Patent Document 2 describes the flow matching conditions and processing for OpenFlow, which is a typical technology for realizing SDN, in units of “flows” defined by combining MAC addresses, IP addresses, and port numbers. A technique for controlling a communication path by setting a method as a flow entry in a controller or a switch is disclosed.

  In OpenFlow, when a switch receives a packet, it searches for a flow entry corresponding to the packet (match), and processes the packet based on the processing method (action) of the matched flow entry. Therefore, OpenFlow is expected to have the advantage of enabling flexible traffic control for network configuration changes and host additions.

  In Patent Document 1, in an OpenFlow network, a communication path is set so as to comprehensively pass through all links on the network, and then a test packet is sent from one measuring terminal and returned from each switch. A technique for monitoring the communication quality (loss rate, average delay, etc.) of each link by observing the behavior of the incoming test packet has been proposed.

Japanese Patent Application 2013-064413

Open Networking Foundation (URL) http://www.openflow.org/ "Latest Trends in Openflow Technology Realizing Next Generation Network 2013" Impress R & D Internet Media Research Institute N. Varris, and J. Manner, "Performance of a Software Switch," in Proc.of HPSR 2011, July 2011. A. Bianco, R. Birke, L. Giraudo, and M. Palacin, "OpenFlow Switching: Data Plane Performance," in Proc. Of ICC, page 1-5, IEEE 2010.

  In the conventional network active measurement technology using OpenFlow, the communication path of the test packet passing through the switch is registered as a flow entry for each switch. Therefore, as the monitoring target network becomes larger, the number of flow entries registered in each switch becomes enormous.

  However, the number of entries that can be registered in each switch is limited due to restrictions such as memory size. Further, the search processing speed decreases as the number of flow entries increases. Therefore, the number of flow entries registered in each switch needs to be as small as possible.

  An object of the present invention is to solve the problems of the prior art and provide a link quality measuring device capable of reducing the number of flow entries of measurement packets registered in each switch on the network, and its flow entry aggregation method and program. is there.

  In order to achieve the above object, the present invention provides a link quality measuring apparatus for relaying a test packet sent from a measuring terminal according to a flow entry registered in each switch on the network, and a flow entry aggregation method and program thereof Is characterized by the following measures.

  (1) In the link quality measuring device of the present invention, a switch in which only a relay switch is connected to each interface link is set with a different identifier for each link, and a relay switch and a terminal switch are connected to each interface link. In (SW10), different identifiers are set for each link connected to the relay switch and a group of links connected to the end switch, and the number of flow entries of each switch is set to each link on the downstream side of each interface of the switch. The number of registered identifiers is aggregated, and in the flow entry of each switch, the correspondence relationship between the identifier registered in each link on the downstream side of each interface and the interface is registered, and each switch is described with an identifier. Sent test packets from the corresponding interface based on the flow entry. It is characterized by being put out.

  (2) The flow entry aggregation method of the link quality measurement apparatus of the present invention includes a first procedure that focuses on a switch having the largest number of flow entries, a second procedure that focuses on an interface having the largest number of flows in the focused switch, A third procedure for aggregating the flow entries related to the interface into one, a fourth procedure for updating and setting a unique identifier for the link through which each flow of the aggregated flow entry passes, and the aggregation result indicating the number of flow entries of the switch of interest A fifth procedure for recalculating based on the above, a sixth procedure for recalculating the number of flow entries of the switch connected to the upstream side of the target switch and its interface based on the aggregation result, and an unfocused switch and interface And a seventh procedure that repeats the first to sixth procedures until there is no more, The entry, characterized in that the correspondence between the identifier and the interface registered in the link downstream of each interface is registered.

  (3) The flow entry aggregation program of the link quality measuring apparatus of the present invention includes a first procedure that focuses on a switch having the largest number of flow entries, a second procedure that focuses on an interface having the largest number of flows in the focused switch, A third procedure for aggregating the flow entries related to the interface into one, a fourth procedure for updating and setting a unique identifier for the link through which each flow of the aggregated flow entry passes, and the aggregation result indicating the number of flow entries of the switch of interest A fifth procedure for recalculating based on the above, a sixth procedure for recalculating the number of flow entries of the switch connected to the upstream side of the target switch and its interface based on the aggregation result, and an unfocused switch and interface The seventh procedure for repeating the first to sixth procedures until there is no more, and the flow of each switch The entry, and an eighth procedure for registering a correspondence between the identifier and the interface registered in the link downstream of each interface, and characterized by causing a computer to execute.

  According to the present invention, the following effects are achieved.

  (1) In SDN such as OpenFlow, when measuring the communication quality of all links with one measuring terminal, each switch can control the packet with a small number of flow entries and measure the communication quality.

  (2) By registering with the number of flow entries minimized, the matching search processing time and memory size used can be reduced and the number of entries for normal data transfer can be increased. It can be used effectively for control.

It is the block diagram which showed the structure of the principal part of the link quality measuring device system with which this invention is applied. It is a figure for demonstrating the aggregation method of a flow entry. It is the figure which showed the example of the flow entry registered into a switch. It is the figure which showed an example of an attention IF management table and a switch aggregation completion table. It is the flowchart which showed the procedure of aggregation processing of the flow entry It is the figure (the 1) which showed typically a mode that the number of flow entries was collected heuristically. It is the figure (the 2) which showed typically a mode that the number of flow entries was collected heuristically. FIG. 10 is a diagram (part 3) schematically illustrating how the number of flow entries is heuristically aggregated. FIG. 6 is a diagram (part 4) schematically illustrating how the number of flow entries is heuristically aggregated. FIG. 10 is a diagram (part 5) schematically illustrating how the number of flow entries is heuristically aggregated. FIG. 10 is a diagram (part 6) schematically illustrating how the number of flow entries is heuristically aggregated. It is the figure which showed a mode that aggregation of a flow entry progressed by repeating an aggregation process. It is the figure which showed typically the mode after the completion of aggregation of a flow entry. It is the figure which showed the method of further reducing the number of flow entries using the priority of OpenFlow. It is the flowchart which showed the procedure which recalculates the aggregation number of a flow entry in the upstream of a focused switch. It is the figure which showed typically the method of recalculating the number of aggregation of a flow entry in the upstream of a focused switch.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. FIG. 1 is a block diagram showing a configuration of a main part of a link quality measuring apparatus system to which the present invention is applied, and is composed of a link quality measuring unit 10 and an SDN network 20 according to an embodiment of the present invention. . Here, the case where the SDN is OpenFlow will be described as an example, and illustrations of components unnecessary for the description of the present invention are omitted.

  The link quality measurement system 10 includes a server device 1 and a measurement terminal 2, and the server device 1 includes a measurement route calculation unit 101, a flow entry aggregation unit 102, and a communication route setting unit 103.

  The network 20 includes an OpenFlow controller 201, a plurality of OpenFlow switches (corresponding to routers), and a plurality of hosts. The controller 201 and each switch communicate with each other using the OpenFlow protocol, each switch is connected by SDN, and data (packet) transfer is performed between each switch and the host.

  In the server device 1, the measurement path calculation unit 101 calculates a test packet communication path by, for example, the method disclosed in Patent Document 1 and acquires a group of test packet communication paths. As will be described in detail later, the flow entry aggregating unit 102 aggregates the flow entries of test packets registered in each switch according to the calculation result of the communication path and passing through the switch to reduce the number of flow entries. . The communication path setting unit 103 registers the aggregated flow entry from the controller 201 to each switch.

  In the measurement terminal 2, the measurement execution unit 104 transmits and receives a test packet, and executes measurement processing such as estimating communication quality of each link from observation data. Here, in this embodiment, it is assumed that the topology information indicating the connection relationship between physical network devices and the management information (for example, MAC address) of each switch are known, and these information are stored in the server device 1. Yes.

  FIG. 2 is a diagram for explaining a flow entry aggregating method by the flow entry aggregating unit 102, and FIG. 3 shows a state in which the flow entries are aggregated by focusing on the switch SW1 in FIG. FIG.

  In FIG. 2, all the packets transmitted from the six interfaces (outgoing links) of the measurement terminal 2 pass through the switch SW0 and are received by the interface eth1 of the switch SW1. In the switch SW1, all test packets destined for the switches SW2, SW3 and SW4 are transmitted from the interface eth2, the test packet destined for the switch SW5 is transmitted from the interface eth3, and the test packet destined for the switch SW6 is an interface. Sent from eth4.

  FIG. 3 is a diagram showing an example of a flow entry registered in the switch SW1, and FIG. 3A is a diagram showing an example of the case where flow entries are not aggregated by the flow entry aggregation unit 102. And one entry for the test packet (destination arrival packet) destined for itself (SW1) and each test packet destined for the five downstream switches SW2, SW3, SW4, SW5 and SW6 ( 11 entries including 5 entries targeted for relay packets) and 5 entries targeted for each test packet (wrapped packet) folded by five downstream switches SW2, SW3, SW4, SW5 and SW6 An entry is registered.

  However, since all the return packets are transmitted from the interface eth1, these entries can be aggregated into one. As a result, as shown in FIG. 3B, the flow entry of the switch SW1 can be reduced to seven.

  Further, all relay packets destined for the switches SW2, SW3, and SW4 are also transmitted from the interface eth2, so that these entries can be aggregated into one. As a result, as shown in FIG. 3C, the flow entry of the switch SW1 can be reduced to five.

  At this time, in order to be able to process a relay packet destined for the three switches SW2, SW3, and SW4 with one entry, the present invention newly introduces an identification concept of “source port (src_port) number”, for example. The src_port number is added in advance to the header field of the packets aggregated in the forward path in each switch SW, and the corresponding switch SW determines the processing by referring to the src_port number.

  Next, the operation of one embodiment of the present invention will be described with reference to a flowchart. FIG. 5 is a flowchart showing a procedure of flow entry aggregation processing according to the present invention. FIGS. 6 to 13 schematically illustrate how the number of flow entries in each switch is heuristically aggregated by repeating the aggregation processing. FIG.

  In the present invention, the number of flow entries is basically reduced by aggregating a plurality of flow entries into one. However, if a plurality of entries having different destination addresses (port numbers) are aggregated, the identification of each entry is impaired. Therefore, the identification concept of “group” (corresponding to the src_port number in FIG. 3) is newly introduced.

  If the identification by destination address or port number is lost due to aggregation, a method is adopted to classify each entry into a different group and ensure the identification of each flow entry by the combination of group identifier Gx and destination address. Has been.

  Here, the initial value of the number of flow entries registered in each switch SW on the network is as shown in FIG. 6, for example, “143” for switch SW1, “100” for switch SW2, and “4” for switch SW3. ... and the explanation starts with all links belonging to the group G1.

[First time: Pay attention to interfaces e1 and 2 of switch SW1]
In step S31, as shown in FIG. 7, the switch with the largest number of flow entries is selected from among the switches that have not yet been aggregated as the current target switch. Here, since all the switches have not yet been aggregated, the switch SW1 having the maximum number of flow entries “143” in all the switches is selected as the current target switch.

  In step S32, the current target switch SW1 selects the interface with the largest number of flow entries from the unfocused interfaces ei, j as the current target interface. The target switch SW1 includes three non-target interfaces e1,2, e1,3, e1,4, and the numbers of flow entries f1,2, f1,3, f1,4 are “101”, “5”, “37”, respectively. Therefore, here, the interfaces e1 and 2 (hereinafter, the switch identifier i may be omitted and simply expressed as e2) are selected as the current interface of interest.

  In step S33, it is determined whether or not the flow entries can be aggregated for the current interface of interest e2. If there are a plurality of flow entries, the process proceeds to step S34 to aggregate these flow entries into one. In step S34, the number of flow entries of the target interface e2 is aggregated from “101” to “1”.

  In step S35, the number of flow entries of the target switch SW1 is recalculated to reflect the aggregation result. In this embodiment, since the number of flow entries of the target interface e2 is reduced from “101” to “1”, the number of flow entries of the target switch SW1 is similarly reduced from “143” to “43”.

  In step S36, all links through which each flow of the aggregated flow entry passes are identified and a unique group ID is updated and registered. In this embodiment, as shown in FIG. 7, the group IDs of the 100 downstream links l1 of the switch SW2 and the link l2 between the switches SW1 and SW2 are updated and registered from the initial values G1 to G2. In step S37, the number of flow entries of all the switches SW upstream from the target switch SW is recalculated.

  FIG. 15 is a flowchart showing a method for recalculating the number of flow entries in step S37. In step S61, it is determined whether or not the target switch is the root SW. Here, since the target switch SW1 is the root SW, this process is skipped. Each process of steps S62 to S64 will be described later.

  Returning to FIG. 5, in step S38, as shown in FIG. 4 (a), “1” is set to the current target interface e2 in the target IF management table to be “focused”. In the target IF management table, “-1” is registered in advance in the interface to which a packet is input, so that it is excluded from the target of attention.

  In step S39, it is determined whether or not all the interfaces ej have been focused for the current focused switch SW1, and since it is determined here that there is an unfocused interface, the process returns to step S31.

[Second time: Focus on SW2 and e2,3]
In the second step S31, as shown in FIG. 8, since the number of flow entries of the switch SW1 is reduced to “43”, the switch SW2 having the number of flow entries “100” is selected as the current switch SW2. Is done.

  In step S32, the current target interface is selected by the target switch SW2. However, since 100 flows are output from each of the 100 non-target interfaces e3, e4,. The description will be continued assuming that the interface e3 is selected as the interface of interest.

  In step S33, it is determined whether or not the flow entry of the target interface e3 can be aggregated. Here, since it is determined that aggregation is not possible, the process proceeds to step S37. In step S37, the number of flow entries is recalculated for all the upstream switches and their corresponding interfaces according to the flowchart of FIG. FIG. 16 is a diagram schematically illustrating a method for recalculating the number of aggregated flow entries on the upstream side of the target switch.

  In FIG. 15, in step S61, since it is determined that the target switch SW2 is not the root SW, the process proceeds to step S62, and the lower limit value of the aggregation number is determined based on the connection status of each switch connected downstream of the target switch SW2. K is calculated.

  In this embodiment, as shown in FIG. 16, it is assumed that all end switches can be integrated into one (for example, group G1) regardless of the number of connections, and one relay switch per switch ( For example, it can be aggregated into groups G2, G3, G4), and the sum of the presence / absence of terminal switches (existing = 1: absent = 0) and the number of relay switches is set as the aggregation number lower limit K. Therefore, the lower limit K of the aggregation number is set to “1” for the target switch SW2.

  In step S63, the aggregation number lower limit K is compared with the current aggregation number (= 1) of the interface e2 of the switch SW1, and it is determined here that the current aggregation number is not larger than the aggregation number lower limit K. Processing is skipped.

  Returning to FIG. 5, in step S38, “1” is set to the current focused interface e3 in the focused IF management table, and the focused IF is set. In step S39, it is determined whether or not all the interfaces have been focused on for the current focused switch SW2, and since there is an unfocused interface here, the process returns to step S31.

[3rd to 101st times: pay attention to SW2, e4, e5, e6 ... e102 sequentially]
Similarly, until the above processing is completed with respect to all the interfaces e2, 3, e2, 4... E2, 102 of the target switch SW2, the switch SW2 continues to be selected as the target switch SW, and each process of steps S31 to S39 Is repeated, and if it is determined that all the interfaces of the target switch SW2 have been focused in the 101st process, the process proceeds to step S40.

  In step S40, in the switch aggregation completion table shown in FIG. 4B as an example, the current focus switch SW2 is marked as focused. In step S41, it is determined whether or not all the switches have been focused. Here, since it is determined that there is an unfocused switch, the process returns to step S31.

[102th: Focus on SW1 and e1,4]
In step S31, as shown in FIG. 9, the switch SW2 has already been focused, and the number of flow entries of the switch SW1 is the maximum value “43”, so that it is selected again as the focused switch SW. In step S32, in the target switch SW1, the interface e4 having the maximum number of flow entries “37” among the untargeted interfaces is selected as the current target interface.

  In step S33, it is determined whether or not aggregation is necessary for the flow entry of the interface of interest e4. Here, since it is determined that it is necessary, the process proceeds to step S34. In step S34, the number of flow entries of the target interface e4 is reduced (aggregated) from “37” to “1”. In step S35, the number of flow entries of the target switch SW1 is recalculated.

  In the present embodiment, since the number of flow entries of the target interface e4 is reduced from “37” to “1”, the number of flow entries of the target switch SW1 is also reduced from “43” to “7”. In step S36, all the links through which each flow of the aggregated flow entry passes (here, all the downstream links of the interface of interest e4) are reorganized into the group G3.

  In step S37, since the target switch SW1 is determined to be the root SW, this process is skipped. In step S38, “1” is set in the target interface e4 in the target IF management table to set the target IF. In step S39, it is determined whether or not all the interfaces have been focused on for the current focused switch SW1. Here, since it is determined that there is an unfocused interface, the process returns to step S31.

[103rd: Focus on SW4 and e4,2]
In step S31, as shown in FIG. 10, the number of flow entries of the switch SW4 becomes the maximum value “36” and is selected as the target switch SW. In step S32, the interface e2 is selected as the current target interface in the target switch SW4.

  In step S33, it is determined whether or not aggregation is necessary for the flow entry of the target interface e2, and since it is determined that it is necessary here, the process proceeds to step S34. In step S34, the number of flow entries of the target interface e2 is reduced (aggregated) from “36” to “1”.

  In step S35, the flow entry number F1 of the target switch SW4 is recalculated and becomes “1”. In step S36, G4 is assigned as a new group ID to all the links through which the flow entry passing through the interface of interest e2 passes.

  In step S37, as the number of flow entries of the target switch SW4 is changed by “1”, the number of flow entries of the switch SW1 upstream of the switch SW is recalculated. Here, since the number of flow entries after aggregation of the target switch SW4 is “1”, the result of recalculation is “1”.

  In step S38, “1” is set in the focused interface e2 in the focused IF management table, and the focused IF is set. In step S39, it is determined whether or not all the interfaces have been focused with respect to the current focused switch SW4. Here, since it is determined that the focus has been focused, the process proceeds to step S40. In step S40, the current focus switch SW4 is regarded as focused. In step S41, it is determined whether or not all the switches have been focused. Here, since it is determined that there is an unfocused switch, the process returns to step S31.

[104th: Focus on SW6 and e6,3]
In step S31, as shown in FIG. 11, since the number of flow entries of the switch SW6 is “35” which is the maximum value, it is selected as the target switch SW. In step S32, the numbers of flow entries of the interfaces e2 and e3 of the target switch SW6 are compared. Here, the interface e3 having the number of interfaces “23” is selected as the target interface.

  In step S33, it is determined whether or not aggregation is necessary for the flow entry of the interface of interest e3. Here, since it is determined that it is necessary, the process proceeds to step S34. In step S34, the number of flow entries of the target interface e3 is reduced (aggregated) from “23” to “1”. In step S35, the number of flow entries of the target switch SW6 is recalculated and reduced from “35” to “13”.

  In step S36, all the links through which the flow entry that passes through the interface of interest e3 passes are extracted and reorganized into the group G5. In step S37, as the number of flow entries of the target switch SW6 is reduced to “13”, the number of flow entries of the switches SW1 and SW4 on the upstream side of the switch SW6 is recalculated.

  In this embodiment, since the number of flow entries of the target switch SW6 is “13”, the number of flow entries of the switch SW4 is also changed to “13”. Similarly, the flow entry number of the switch SW1 is also changed to “13”.

  In step S38, “1” is set in the current focused interface e3 in the focused IF management table, and the focused IF is set. In step S39, it is determined whether or not all the interfaces have been focused on the current focused switch SW6. Here, since there is an unfocused interface e2, the process returns to step S31.

  This process is repeated until it is determined in step S41 that all switches have been consolidated. In step S42, the group identifiers are reassigned in ascending order.

  FIG. 12 is a diagram showing how the aggregation of the flow entries in each switch proceeds heuristically by repeating the above aggregation processing, and FIG. 13 shows the completion of the aggregation of the flow entries and the reassignment of the group identifier. It is the figure which showed the state.

  In a switch (SW1, SW3, SW4, SW6, SW7, SW8) in which only the relay switch is connected to the link of each interface, a unique identifier is set for each link. In the switch (SW10) in which the relay switch and the end switch are connected to the link of each interface, a unique identifier is set for each link connected to the relay switch and a group of links connected to the end switch.

  The number of flow entries of each switch is aggregated to the number of identifiers registered in each link on the downstream side of each interface of the switch, and the flow entry of each switch is registered in each link on the downstream side of each interface. The correspondence between the identifier and the interface is registered, and each switch sends out a test packet in which the identifier is described from the corresponding interface based on the flow entry.

  On the other hand, from the server device to the measurement terminal 2, registration of the src_port number corresponding to each group ID is requested as an identifier to be described in each test packet. In this embodiment, for each test packet, registration of the identifier registered in the incoming link of the destination switch is requested.

  For example, an additional registration of src_port number = 106 corresponding to the group G6 is requested to the test packet addressed to each switch connected to the outgoing link of the switch SW13. Similarly, an additional registration of src_port number = 105 corresponding to group G5 is requested for the test packet addressed to each switch connected to the outgoing link of switch SW9, and the address addressed to each switch connected to the outgoing link of switch SW5 is requested. The test packet is requested to be additionally registered with src_port number = 107 corresponding to group G7. In response to the additional registration request for the identifier, the measuring terminal 2 adds each requested src_port number to, for example, the header field of the test packet, and then transmits it to the network.

  In addition, as shown in FIG. 14, by using priority items that can be specified in OpenFlow and assigning priorities, items other than the previously specified items are set to “other than the above” and nothing is specified in the match condition. Can further reduce the number of flow entries.

  According to the present embodiment, in SDN such as OpenFlow, the communication quality of all links can be measured with one measuring terminal while keeping the number of flow entries of each switch small. In addition, by minimizing and registering the number of flow entries, the matching search processing time can be reduced and the memory size used can be reduced, and the number of entries for normal data transfer can be increased. Quality measurement that prioritizes certain data transfer becomes possible.

  DESCRIPTION OF SYMBOLS 1 ... Server apparatus, 2 ... Measurement terminal, 10 ... Link quality measurement system, 20 ... Network, 101 ... Measurement path calculation part, 102 ... Flow entry aggregation part, 103 ... Communication path setting part, 201 ... OpenFlow controller

Claims (6)

In the link quality measurement device that relays the test packet sent from the measurement terminal according to the flow entry registered in each switch on the network,
In the switch where only the relay switch is connected to the link of each interface, a different identifier is set for each link.
In a switch in which a relay switch and a terminal switch are connected to the link of each interface, different identifiers are set for each link connected to the relay switch and a group of links connected to the terminal switch,
The number of flow entries of each switch is aggregated to the number of identifiers registered in each link on the downstream side of each interface of the switch,
In the flow entry of each switch, the identifier registered in each link on the downstream side of each interface and the corresponding relationship with the interface are registered.
Each switch sends out a test packet in which an identifier is described from a corresponding interface based on the flow entry. Means for calculating a measurement path of the test packet;
Means for aggregating the flow entries of each switch corresponding to the measurement path;
Means for registering each aggregated flow entry with a corresponding switch,
The means for aggregating is
A first procedure focusing on the switch with the largest number of flow entries;
A second procedure focusing on the interface with the maximum number of flows in the target switch;
A third procedure for consolidating the flow entries related to the interface of interest into one;
A fourth procedure for updating and registering a unique identifier for a link through which each flow of the aggregated flow entry passes;
A fifth procedure for recalculating the number of flow entries of the target switch based on the aggregation result;
A sixth procedure for recalculating the number of flow entries of the switch connected to the upstream side of the target switch and its interface based on the aggregation result;
The link quality measuring apparatus according to claim 1, wherein a seventh procedure that repeats the first to sixth procedures is executed until there are no unfocused switches and interfaces.   The link quality measuring apparatus according to claim 2, wherein an identifier registered in an incoming link of a destination switch is registered in each test packet.   The link quality measuring apparatus according to claim 2 or 3, wherein in the aggregation step, the flow entries of the return path test packets returned to the measurement terminal are further consolidated into one. A method in which a computer aggregates the entry flow of test packets registered in each switch on a network,
A first procedure focusing on the switch with the largest number of flow entries;
A second procedure focusing on the interface with the maximum number of flows in the target switch;
A third procedure for consolidating the flow entries related to the interface of interest into one;
A fourth procedure for updating and registering a unique identifier for a link through which each flow of the aggregated flow entry passes;
A fifth procedure for recalculating the number of flow entries of the target switch based on the aggregation result;
A sixth procedure for recalculating the number of flow entries of the switch connected to the upstream side of the target switch and its interface based on the aggregation result;
And a seventh procedure that repeats the first to sixth procedures until there are no unfocused switches and interfaces,
A flow entry aggregation method for a link quality measuring apparatus, characterized in that an identifier registered in each link on the downstream side of each interface and a corresponding relationship with the interface are registered in the flow entry of each switch. A program that causes a computer to aggregate flow entries of test packets registered in each switch on the network,
A first procedure focusing on the switch with the largest number of flow entries;
A second procedure focusing on the interface with the maximum number of flows in the target switch;
A third procedure for consolidating the flow entries related to the interface of interest into one;
A fourth procedure for updating and registering a unique identifier for a link through which each flow of the aggregated flow entry passes;
A fifth procedure for recalculating the number of flow entries of the target switch based on the aggregation result;
A sixth procedure for recalculating the number of flow entries of the switch connected to the upstream side of the target switch and its interface based on the aggregation result;
A seventh procedure for repeating the first to sixth procedures until there are no unfocused switches and interfaces;
A flow entry aggregation program for a link quality measurement apparatus that causes a computer to execute an eighth procedure for registering an identifier registered in each link on the downstream side of each interface and a corresponding relationship with the interface in a flow entry of each switch.