JP2012039529A - Power characteristic use communication network setting method and network device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce power consumed by a network by selecting a communication path of high power efficiency.SOLUTION: A power characteristic use communication network setting method comprises the steps of: acquiring flow information and load information from each router in a condition that a network includes routers with links being capable of RA and links being incapable of RA mixed; storing them in first storage means; calculating a path according to traffic condition based on the flow information and the load information of the first storage means, power characteristic information, and network configuration information; referring, based on the calculated path, to second storage means storing power characteristic information of a line card configuring both ends of each link and configuration information of the network; accessing each router; and changing the setting.

Description

  The present invention relates to a power characteristic use communication network setting method and a network device, and more particularly, power for setting a communication path for reducing power consumption of a router used in a managed IP (Internet Protocol) relay network. The present invention relates to a characteristic use communication network setting method and a network device.

  In routers, unlike servers, power consumption in the standby state does not differ significantly compared to the maximum load (the standby power is 90% or more of the maximum load in large cases), so the standby power can be greatly reduced. Has been proposed to utilize the sleeve state (see, for example, Non-Patent Document 1).

  However, in recent years, rate-adaption (RA) technology that suppresses power consumption by reducing the voltage and operating frequency of computing devices at low loads has been advancing in router bodies and line cards (for example, non-patented) Reference 2).

Joseph Chabarek, Joel Sommers, Paul Barford, et. Al "Power Awareness in Network Design and Routing" In Proc. IEEE INFOCOM, 2008, http://pages.cs.wisc.edu/~pb/infocom08a_final.pdf Sergiu Nedevschi, Lucian Popa, Gianluca lannaccone, et. Al. "Reducing Network Energy consumpition via Sleeping and Rate-adaption" Technical Report No. UCB / EECS-2007-128. Http://www.eecs.berkeley.edu/Pubs /TechRpts/2007/EECS-2007-128.pdf

  Most of the currently proposed methods are based on the premise that all devices in the network are capable or not capable of RA. However, in actuality, it is thought that the transition will be made through a mixture of devices that are impossible to listen to in the future and where RA is possible.

  Methods that take this hybrid state into account (for example, reference 1 “Juan camilo Cardona Restrepo, Claus G. Gruber, Carmen Mas Machuca,“ Energy Profile Aware Routing ”, ICC conf. Proc, 2009. http://ieeexplore.ieee. org / stamp / stamp.jsp? arnumber = 05208041 & tag = 1 ”), but there is a possibility that an appropriate measure cannot be taken when the load fluctuates because the load that actually flows through the network is not taken into consideration.

  The present invention has been made in view of the above points, and provides a power characteristic use communication network setting method and a network device capable of reducing the power consumed by the network by selecting a power efficient communication path. For the purpose.

In order to solve the above-described problems, the present invention (Claim 1) reduces the power consumption of a router in a network in which a router link can suppress power consumption and cannot. A monitoring and control device for
It consists of router information collection means, router control means, route calculation means, router control means,
The router information collecting means includes
It has means for acquiring flow information and load information from each router and holding it in the first storage means,
The router control means includes
A second storage means for holding the power characteristic information of the line cards constituting both ends of each link and the network configuration information; and the second storage means based on the route calculated by the route calculation means. Have means to browse, access each router and change settings,
The route calculation means includes
Means for calculating a route according to a traffic situation based on the flow information of the first storage means, the load information, the power characteristic information of the second storage means, and the configuration information of the network. It is characterized by.

Further, the present invention (Claim 2) is the route calculation means according to Claim 1,
A third storage means for holding the costs of all links at the present time;
Traffic matrix generation means for creating a matrix of traffic flowing in the network from the flow information stored in the first storage means;
Based on the load information stored in the first storage means, if there is no link whose link utilization rate exceeds a predetermined threshold, the link with the lowest load is based on a predetermined priority based on power characteristics. Utilization rate judging means for increasing the cost;
Link load calculating means for performing a simulation based on the traffic matrix and the configuration information of the network stored in the second storage means, and calculating a utilization factor of each link;
And minimum load link cost changing means for increasing the value of the cost of the minimum load link on the simulation model based on the calculated link utilization rate and the power characteristic of each link.

Further, the present invention (Claim 3) is the utilization rate determination means of Claim 2,
The priority is set in the order of on / off type, mixed type, and cubic power characteristics, and a sleep flag is set to link off to a link having a minimum utilization rate less than a predetermined threshold, and the link having the sleep flag set Including means to increase the cost of traffic and prevent traffic from flowing,
In the link load calculation means,
With respect to the utilization rate of each link obtained by the simulation, there is a link having a utilization rate exceeding the predetermined threshold, or the link between the edge routers is not lost by the link having the Sleep flag set. If there is a link with a utilization rate exceeding the threshold value, or communication is lost due to the link having the Sleep flag set, the link cost not exceeding the threshold value is obtained from the third storage means. A link cost reference means for selecting a link to have,
Link list generating means for extracting, listing, and passing to the router control means links that are not loaded based on the link utilization rate selected by the link cost reference means.

Further, according to the present invention (Claim 4), in a network in which router links are both capable of suppressing power consumption and impossible, power characteristic utilization communication for reducing power consumption of the router is provided. A network setting method,
In an apparatus comprising router information collection means, router control means, route calculation means, router control means,
The router information collecting means obtains flow information and load information from each router, and retains it in the first storage means;
A route calculation step in which the route calculation unit calculates a route according to a traffic situation based on the flow information, the load information, and power characteristic information and network configuration information of the first storage unit;
The router control means refers to the second storage means for holding the power characteristic information of the line card constituting the both ends of each link and the network configuration information based on the route calculated by the route computing means, A router control step to access the router and change the configuration;
It is characterized by performing.

Further, the present invention (Claim 5) is the route calculation step of Claim 4,
Holding the cost of all links at the current time in a third storage means;
A traffic matrix generation step of creating a matrix of traffic flowing through the network from the flow information stored in the first storage means;
Based on the load information stored in the first storage means, if there is no link whose link utilization rate exceeds a predetermined threshold, the link with the lowest load is based on a predetermined priority based on power characteristics. A utilization rate determining step that increases costs;
A link load calculating step of performing a simulation based on the traffic matrix and the configuration information of the network stored in the second storage means, and calculating a utilization rate of each link;
And a minimum load link cost changing step for increasing the cost value of the minimum load link on the simulation model based on the calculated link utilization rate and the power characteristic of each link.

Moreover, this invention (Claim 6) is the utilization rate determination step of Claim 5,
The priority is set in the order of on / off type, mixed type, and cubic power characteristics, and a sleep flag is set to link off to a link having a minimum utilization rate less than a predetermined threshold, and the link having the sleep flag set Increasing the cost of traffic and preventing traffic from flowing,
In the link load calculation step,
With respect to the utilization rate of each link obtained by the simulation, there is a link having a utilization rate exceeding the predetermined threshold, or the link between the edge routers is not lost by the link having the Sleep flag set. If there is a link with a utilization rate exceeding the threshold value, or communication is lost due to the link having the Sleep flag set, the link cost not exceeding the threshold value is obtained from the third storage means. A link cost reference step for selecting a link having,
A link list generating step of extracting, listing, and passing to the router control means links that are not loaded based on the link utilization rate selected in the link cost reference step.

  The present invention (Claim 7) is a power characteristic utilization communication network setting program for causing a computer to function as each means constituting the monitoring control apparatus according to any one of Claims 1 to 3. .

  As described above, according to the present invention, it is possible to simply sleep by selecting a power-efficient communication path that matches the traffic flow that flows in a power-efficient crisis and a mixture of bad devices that will occur in the future. It becomes possible to reduce electric power rather than introducing it.

  In addition, since the route is selected in accordance with the characteristics of the power consumption with respect to the link utilization efficiency of the network device and the route selection policy, it can be expected to follow up when the traffic state changes slowly, such as at night or during the daytime.

It is a block diagram of the network of this invention. It is a power characteristic of both types of line cards with respect to link utilization. It is a power consumption characteristic (EP) of each link type with respect to a link utilization factor. It is a flowchart of the network operation | movement in one embodiment of this invention. It is a block diagram of the monitoring control apparatus in one embodiment of this invention. It is a flowchart of the process of the monitoring control apparatus in one embodiment of this invention.

  Embodiments of the present invention will be described below with reference to the drawings.

  First, the preconditions of the present invention will be described.

  1. About half of the power consumed by the router is the chassis, and most of the rest is made up of line cards. The power consumption of the line card increases as the processing speed increases (Non-Patent Document 1, Reference 2 “Bart Puype, Willem Vereechen, Didier Colle, et.al.,“ Power Reduction Techiques in Multiayer Traffic engineering ”, ICTON 2009. ").

  2. There are the following two types of power consumption in router line cards.

a: Since there is almost no difference between standby power and power consumption at maximum load (see Non-Patent Document 1), it is necessary to introduce sleep in order to reduce power consumption;
b: Utilizing Rate Adaption (RA) technology (see Non-Patent Document 2) that reduces standby power by lowering the voltage and operating frequency of an arithmetic device in a light load state;
3. By using a traffic flow measurement technology such as NetFlow or sFlow, it is possible to collect information on flows that flow through each device to a specific device (flow collector) (Reference 3: Cisco NetFlow v3, http://www.cisco). .com / warp / public / cc / pd / nemnsw / nesv / prodlit / 2033_pp.pdf, reference 4: sFlow, see http: //sflow.org.about.index.php).

  Next, a network targeted by the present invention will be described.

  FIG. 1 shows a network configuration to which the present invention is applied.

  A network 100 shown in the figure is a network having AS (Autonomous System) border routers having the same routing policy, and has a redundant configuration.

  The line card of each router 10 has an RA capable and an RA impossible, and notifies the monitoring control device 20 of traffic flow information flowing through itself.

  The monitoring control device 20 connected to the network 100 manages the power characteristic (Energy Profile: EP) information of the line cards that constitute both ends of each link and the configuration information of the network. Also, it functions as a flow collector and acquires flow information and load information from each router 10. Each router can be accessed via Telnet or SSH (Secure Shell) to change the settings.

There are some line cards that can use Rate Adaption (RA) technology that varies the voltage and operating frequency according to the load situation of the router itself, and some that can not, and each has different power consumption characteristics for the link load . Usually, the maximum power consumption of an arithmetic unit has a positive correlation with its processing speed, so that RA-compatible and non-RA-compatible models with similar processing speeds have lower power consumption. Note that the RA-compatible type consumes less power.
-There is a positive correlation between the maximum power consumption of the computing device and the processing speed;
RA technology does not reduce maximum power consumption;
That is why.

  FIG. 2 shows the power characteristics of both types of line cards versus link utilization.

  As shown in FIG. 6A, the power consumption of the non-RA-compliant type is almost dependent on the on / off state, and in the on state, a value close to the maximum power consumption is consumed regardless of the link utilization rate (non-patented). Reference 1).

  Also, as shown in FIG. 5B, the power consumption of the RA-compatible arithmetic unit is the product of the square of the voltage, the operating frequency, and the capacitance, and therefore the voltage and the operating frequency vary depending on the utilization rate. When the RA technology is used, the power consumption for the load increases to the third power (see Document 1).

  FIG. 3 shows the power consumption characteristics of each link type with respect to the link utilization rate.

  The trend of power consumption with respect to the utilization rate when each link is used is classified into the following power characteristics (EP) according to the combination of RA of line cards at both ends of the link. In the following, this is called the link EP. Since the line cards are connected to both ends of each link, the link EP is a combination of the EPs of both line cards.

(A) On / off type (non-RA type-non-RA type)
FIG. 4A shows the power characteristics of a link in which the line cards at both ends have a large standby power, and the maximum power consumption is consumed regardless of the utilization rate as long as the power is on.

(B) Cubic type (RA type-RA type)
FIG. 6B shows the power characteristics of a link in which the line cards at both ends can be RA, which has a low standby power, and the power consumption increases in a cubic function type according to the load.

(C) Mixed type (RA type-non-RA type)
Figure (c) shows the power characteristics of a link with a line card capable of RA at one end and a line card not capable of RA at the other end. It has higher standby power than Cubic and consumes power according to the load. Increases to a cubic function type.

The power savings of the above three types when Sleep is introduced
On / off type ≧ mixing type ≧ cubic type.

  Based on the above assumptions, the present invention reallocates link costs according to the power consumption characteristics (EP: Energy Profile) and link selection policy based on the flow information observed at each router. To do. The process of redistribution and differential routing is performed by simulation, and this is applied to a real network after a more power efficient link setup is determined.

  Next, a series of operations in the network shown in FIG. 1 will be described.

  FIG. 4 is a flowchart of the network operation according to the embodiment of the present invention.

  Step 101) In an initial state, the network 100 determines a route according to a routing protocol set in advance.

  Step 102) Each router 10 notifies the monitoring control device 20 of load information applied to its own interface by SNMP (Simple Network Management Protocol).

  Step 103) The edge router collects the flow information necessary for creating the traffic matrix and transmits it to the monitoring controller 20 at regular intervals.

  The monitoring control device 20 performs path setting in consideration of power efficiency by the following operation (details will be described later).

  Step 104) The monitoring controller 20 generates a traffic matrix from the obtained flow information.

  Step 105) Here, based on the load information of each link, it is determined whether or not sleep introduction is possible, and if sleep introduction is possible, the process proceeds to step 106, and if impossible, the process proceeds to step 107.

  Step 106) If possible, increase the cost of the link with the lowest load among the links of the prioritized EPs so that traffic does not flow. The purpose is to secure the product).

  Step 107) The monitoring and control apparatus 20 performs a simulation using the network model and traffic matrix whose cost has been reset, and calculates the load on all links.

  Step 108) If there is no link having a higher load than the predetermined value in the calculated link load, the process proceeds to Step 109, and if there is a high load link, the process proceeds to Step 110.

  Step 109) When there is no link whose calculated link load is higher than a predetermined value, the link cost of the lowest load among the links of the prioritized EPs is increased.

  Step 110) When the final link cost and the route determination based on the final link cost are made, the supervisory control device 20 sets the link cost to the real network and introduces the Sleep to the unused link.

  Next, the monitoring control device 20 will be described.

  FIG. 5 shows a configuration of the monitoring control apparatus according to the embodiment of the present invention.

  The monitoring control device 20 includes a communication interface unit 21, a router information collection unit 22, a router control unit 23, a control unit 24, and a route calculation unit 25. Information is exchanged between the constituent elements (arrows in the figure). Monitor and control.

  The communication interface unit 21 has an IP address and performs communication processing with each managed device.

  The router information collection unit 22 includes a storage unit and a control unit. The storage unit is a storage medium such as a memory or a disk, and may be connected inside or outside the router information collection unit 22. The control unit operates as a flow collector or an SNMP manager, accesses a device to be monitored via the communication interface unit 21, collects load information according to the following pattern, and stores it in the storage unit.

  • Regular: Requests load information from each device periodically and collects information.

  Non-periodic: Collects information sent from each device when a change that affects each device or network occurs, such as when the load increases in a power-saving state.

  Specific: Access to a specific router via Telnet or SSH when the state changes, and load information is collected using the Show Interface command.

  Based on the information collected from the router information collection unit 22, the control unit 24 determines the calculation start in the route calculation unit 25 and the link cost setting, and the setting of each router in the router control unit 23, Instruct.

  The router control unit 23 includes a storage unit and a control unit. The storage unit of the router control unit 23 is a storage medium such as a memory or a disk, and may be connected inside or outside the router control unit 23. The control unit of the router control unit 23 performs on / off operations of router components and links and the like for each router to be monitored. The setting information is stored in the storage unit. It also has a list of link information between routers. The link list has line card information at both ends of the link, the EP of the link, a sleep flag, and cost information.

  The route calculation unit 25 starts calculation in response to a command from the control unit 24. Based on the setting information of the router control unit 23 and the flow information collected by the router information collection unit 22, a simulation is performed, the load of each link is calculated, and the cost is reset.

  Below, operation | movement of the monitoring control apparatus 20 in said structure is demonstrated.

  FIG. 6 is a flowchart of processing of the monitoring control apparatus according to the embodiment of the present invention.

  Step 201) Based on the traffic flow information stored in the storage unit of the router information collecting unit 22 via the communication interface unit 21, a matrix of traffic flowing through the network is created. The traffic matrix creation method is described in Reference 5, “A Gunnar, M Johansson, t Telkamp,“ Traffic Matrix Estimation on a Large IP Back bone-A comparison on Real Data ”http://portal.acm.org/citation.cfm?id = 1028788.1028807 "can be used. At this time, in order to ensure the communication between all the edges, a traffic having a value lower than the minimum traffic amount is given to the edge ground where no flow is observed.

  Step 202) The router information collecting unit 22 collects load information of all the links of the network by SNMP and stores it in the storage unit.

  Step 203) Based on the load information stored in the storage unit of the router information collection unit 22, the route calculation unit 25 determines whether there is a link whose utilization rate exceeds the threshold value α. If there is no link exceeding α, the processing from step 204 is performed. If there is a link exceeding α, do nothing and wait.

  Step 204) The route calculation unit 25 stores the current costs of all links in the memory.

  Step 205) In order to remove a link having a low utilization rate but no alternative route, a link information list excluding links set with a Sleep flag, which will be described later, is created and stored in a memory (not shown).

  Step 206) It is determined whether or not there is a link of less than α on / off (P1) for EP in the created link list. If there is a link, the process proceeds to Step 207. If not, the process proceeds to Step 208.

  Step 207) The Sleep flag is set for the minimum utilization link at P1, and the routine proceeds to Step 211.

  Step 208) Next, it is determined whether there is a link of less than α in the mixed type (P3). If there is a link, the process proceeds to Step 209, and if not, the process proceeds to Step 210.

  Step 209) The Sleep flag is set for the minimum utilization link at P3, and the process proceeds to Step 211.

  Step 210) The Sleep flag is set for the minimum utilization link in Cubic (P2), and the process proceeds to Step 211.

  The processing from Step 206 to Step 210 is less than the threshold value α in the link list having the respective profiles with the priority of ON / OFF (P1)> mixed type (P3)> Cubic (P2) for EP in the created link list. The link with the lowest utilization is identified and the Sleep flag is set.

  Step 211) The cost of the link with the Sleep flag set is increased, and traffic is prevented from flowing.

  Step 212) Based on the traffic matrix and network configuration information stored in the storage unit of the router information collection unit 22, a simulation is performed to calculate the utilization rate of each link.

  Step 213) It is determined whether or not there is a link having a utilization rate exceeding the threshold α with respect to the calculated utilization rates of all links. If there is a link, the process proceeds to Step 215, and if not, the process proceeds to Step 214.

  Step 214) It is determined whether or not the communication between the edge routers is lost by the link with the Sleep flag set. If there is communication, the process proceeds to Step 204. Otherwise, the process proceeds to Step 215.

  Step 215) When there is a link exceeding the threshold α (Step 213, Yes), or when the link between the edge routers is lost due to the link with the Sleep flag set (Step 214, No), the threshold is exceeded. The link cost setting information stored in the memory in step 204 having no link cost setting is read and used.

  Step 216) With reference to the utilization rate information of each link calculated from the read link cost setting information, a link with no load (zero load link) is extracted and listed.

  Step 217) The control unit 24 acquires the link cost information from the route calculation unit 25, and applies the link cost information to the actual network.

  Step 218) As a result of the cost setting and routing, the router having the zero load link is accessed from the router control unit 23 by Telnet or SSH, and the interface of the unused link is stopped to enter the sleep state.

  According to the above-described embodiment, assuming that the router links constituting the network have both RA-capable and impossible links, power-efficient communication in accordance with the flowing traffic flow Power consumption is reduced by selecting a route.

  The operation of the components of the monitoring control device 20 shown in FIG. 5 can be constructed as a program, installed in a computer used as the monitoring control device, executed, or distributed via a network.

  In addition, the constructed program can be stored in a portable storage medium such as a hard disk, a flexible disk, or a CD-ROM, and can be installed or distributed in a computer.

  The present invention is not limited to the above-described embodiment, and various modifications and applications can be made within the scope of the claims.

DESCRIPTION OF SYMBOLS 10 Router 20 Monitoring control apparatus 21 Communication interface part 22 Router information collection part 23 Router control part 24 Control part 25 Path | route calculating part 100 Network

Claims (7)

In a network where router links are capable of suppressing power consumption and those where power consumption is not possible, a monitoring and control device for reducing power consumption of the router,
It consists of router information collection means, router control means, route calculation means, router control means,
The router information collecting means includes
It has means for acquiring flow information and load information from each router and holding it in the first storage means,
The router control means includes
A second storage means for holding the power characteristic information of the line cards constituting both ends of each link and the network configuration information; and the second storage means based on the route calculated by the route calculation means. Have means to browse, access each router and change settings,
The route calculation means includes
Means for calculating a route according to a traffic situation based on the flow information of the first storage means, the load information, the power characteristic information of the second storage means, and the configuration information of the network;
A monitoring control device characterized by that. The route calculation means includes
A third storage means for holding the costs of all links at the present time;
Traffic matrix generation means for creating a matrix of traffic flowing in the network from the flow information stored in the first storage means;
Based on the load information stored in the first storage means, if there is no link whose link utilization rate exceeds a predetermined threshold, the link with the lowest load is based on a predetermined priority based on power characteristics. Utilization rate judging means for increasing the cost;
Link load calculating means for performing a simulation based on the traffic matrix and the configuration information of the network stored in the second storage means, and calculating a utilization factor of each link;
A minimum load link cost changing means for increasing the cost value of the minimum load link on the simulation model based on the calculated link utilization rate and the power characteristic of each link;
The monitoring control device according to claim 1, comprising: The utilization rate determining means includes
The priority is set in the order of on / off type, mixed type, and cubic power characteristics, and a sleep flag is set to link off to a link having a minimum utilization rate less than a predetermined threshold, and the link having the sleep flag set Including means to increase the cost of traffic and prevent traffic from flowing,
The link load calculating means includes
With respect to the utilization rate of each link obtained by the simulation, there is a link having a utilization rate exceeding the predetermined threshold, or the link between the edge routers is not lost by the link having the Sleep flag set. If there is a link with a utilization rate exceeding the threshold value, or communication is lost due to the link having the Sleep flag set, the link cost not exceeding the threshold value is obtained from the third storage means. A link cost reference means for selecting a link to have,
Based on the utilization rate of the link selected by the link cost reference means, a link list generating means that extracts and lists links that are not loaded, and passes them to the router control means;
The monitoring control device according to claim 2, comprising: In a network in which router links are both capable of suppressing power consumption and incapable of power consumption, a power characteristic utilization communication network setting method for reducing power consumption of a router,
In an apparatus comprising router information collection means, router control means, route calculation means, router control means,
The router information collecting means obtains flow information and load information from each router, and retains it in the first storage means;
A route calculation step in which the route calculation unit calculates a route according to a traffic situation based on the flow information, the load information, and power characteristic information and network configuration information of the first storage unit;
The router control means refers to the second storage means for holding the power characteristic information of the line card constituting the both ends of each link and the network configuration information based on the route calculated by the route computing means, A router control step to access the router and change the configuration;
A communication network setting method using power characteristics, characterized in that: In the route calculation step,
Holding the cost of all links at the current time in a third storage means;
A traffic matrix generation step of creating a matrix of traffic flowing through the network from the flow information stored in the first storage means;
Based on the load information stored in the first storage means, if there is no link whose link utilization rate exceeds a predetermined threshold, the link with the lowest load is based on a predetermined priority based on power characteristics. A utilization rate determining step that increases costs;
A link load calculating step of performing a simulation based on the traffic matrix and the configuration information of the network stored in the second storage means, and calculating a utilization rate of each link;
A minimum load link cost changing step for increasing the cost value of the minimum load link on the simulation model based on the calculated link utilization and the power characteristics of each link;
The method for setting a communication network using power characteristics according to claim 4. In the utilization rate determining step,
The priority is set in the order of on / off type, mixed type, and cubic power characteristics, and a sleep flag is set to link off to a link having a minimum utilization rate less than a predetermined threshold, and the link having the sleep flag set Increasing the cost of traffic and preventing traffic from flowing,
In the link load calculation step,
With respect to the utilization rate of each link obtained by the simulation, there is a link having a utilization rate exceeding the predetermined threshold, or the link between the edge routers is not lost by the link having the Sleep flag set. If there is a link with a utilization rate exceeding the threshold value, or communication is lost due to the link having the Sleep flag set, the link cost not exceeding the threshold value is obtained from the third storage means. A link cost reference step for selecting a link having,
Based on the link utilization rate selected in the link cost reference step, links that are not loaded are extracted, listed, and a link list generation step that is passed to the router control means;
The method for setting a communication network using power characteristics according to claim 5.   A power characteristic utilization communication network setting program for causing a computer to function as each means constituting the monitoring control device according to any one of claims 1 to 3.

Images