CN109617704B - 40-100Gbps Ethernet energy-saving strategy implementation method based on prediction - Google Patents

40-100Gbps Ethernet energy-saving strategy implementation method based on prediction Download PDF

Info

Publication number
CN109617704B
CN109617704B CN201811582071.1A CN201811582071A CN109617704B CN 109617704 B CN109617704 B CN 109617704B CN 201811582071 A CN201811582071 A CN 201811582071A CN 109617704 B CN109617704 B CN 109617704B
Authority
CN
China
Prior art keywords
energy
state
saving
eee
period
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201811582071.1A
Other languages
Chinese (zh)
Other versions
CN109617704A (en
Inventor
蒋万春
廖凯琴
严瑜龙
王建新
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Central South University
Original Assignee
Central South University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Central South University filed Critical Central South University
Priority to CN201811582071.1A priority Critical patent/CN109617704B/en
Publication of CN109617704A publication Critical patent/CN109617704A/en
Application granted granted Critical
Publication of CN109617704B publication Critical patent/CN109617704B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L12/00—Data switching networks
    • H04L12/02—Details
    • H04L12/12—Arrangements for remote connection or disconnection of substations or of equipment thereof
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/14—Network analysis or design
    • H04L41/142—Network analysis or design using statistical or mathematical methods
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/14—Network analysis or design
    • H04L41/145—Network analysis or design involving simulating, designing, planning or modelling of a network
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/14—Network analysis or design
    • H04L41/147—Network analysis or design for predicting network behaviour
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D30/00—Reducing energy consumption in communication networks
    • Y02D30/50—Reducing energy consumption in communication networks in wire-line communication networks, e.g. low power modes or reduced link rate

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

The patent discloses a 40-100Gbps Ethernet energy-saving strategy implementation method based on prediction. The ieee802.3bj standard defines two Energy-saving states, namely Fast Wake (Fast Wake) and Deep Sleep (Deep Sleep), for 40-100Gbps Energy-saving Ethernet (EEE), and the Energy-saving states and the state transition delays corresponding to different Energy-saving states are different. On the premise of ensuring that the tail delay of the data frame does not exceed the preset expected delay based on the predicted 40-100Gbps Ethernet energy-saving strategy, the arrival number of the data frames in the next period is periodically predicted according to the historical information, and the energy-saving state used in the next period is selected and the time of EEE leaving the energy-saving state is controlled based on the prediction result, so that the better compromise is achieved in the aspects of improving energy-saving capacity and reducing tail delay. Experimental results show that the 40-100Gbps Ethernet energy-saving strategy based on prediction designed by the patent can achieve a better energy-saving effect compared with the existing energy-saving strategy while controlling the tail delay of a data frame within an expected value.

Description

40-100Gbps Ethernet energy-saving strategy implementation method based on prediction
Technical Field
The invention relates to the field of 40-100Gbps energy-saving Ethernet, in particular to a method for implementing 40-100Gbps Ethernet energy-saving strategy based on prediction.
Background
In recent years, with 40-100Gbps Ethernet in commercial use, the IEEE802.3bj working group promulgated the corresponding Energy Efficient Ethernet (EEE) standard in 2014, 9. In this standard, the link power consumption in the deep sleep state is only 10% of the peak and the time required to transition to the active state is 5.5 μ s for the deep sleep state, while the link power consumption in the fast wake-up state is typically 70% of the peak power consumption, but the time required to transition from the fast wake-up state to the normal operating state is shorter, 0.34 μ s. EEE uses both states to save energy. When no data frame needs to be transmitted in the link, the link can be selected to enter one of the low power consumption states, and in the process, the link power consumption is 100%. On the one hand, as the link capacity increases, the power consumption of the ethernet increases considerably, and on the other hand, as the link bandwidth increases, the expected delay of data frames becomes smaller and smaller, while the time required for frequent state transitions causes the frame delay to increase relatively. Therefore, reducing unnecessary state transitions is crucial for power saving. The energy saving of the EEE depends on the energy saving state selected by the link and the time spent in this energy saving state, the longer the dwell time, the more energy is saved. In order to prolong the retention time of the EEE in the power saving state, similarly to the method in the 1-10Gbps power saving ethernet, the data frames arriving during the state transition and in the power saving state are stored, so that the link does not exit from the power saving state immediately. Although this may greatly reduce link power consumption, it may additionally increase the delay of data frames.
Several power saving strategies, Dual-Mode, FC-SSHI and FC-DT, have emerged in 40-100Gbps Ethernet networks. As shown in fig. 2, the Dual-Mode strategy does not need to select the power saving state, and when there is no data frame in the link, the link always directly switches from the working state to the fast wake-up state; if a data frame arrives when the link is in the fast wake-up state, the link is immediately woken up, otherwise, the link stays in the fast wake-up state for TFAfter time TFtoDTo a deep sleep state until a data frame arrives at the link. In FIG. 3, the FC policy is an improvement of the Dual-Mode policy, and the link stays at T in the fast wake-up stateFReaches a threshold value C set by the systemFThen, the link is converted into a deep sleep state; only when the number of accumulated data frames reaches the threshold value C set by the systemDThe link is only woken up from the deep sleep state. The FC-SSHI strategy in FIG. 4 is also improved based on the Dual-Mode strategy. If the number of data frames accumulated in the conversion process of the previous energy-saving state is lower than the system set threshold value CDThe next time the link can be transitioned directly from the working state to the deep sleep state. However, none of the three strategies can control the delay while achieving good energy saving effect. Whereas the FC-DT strategy of fig. 5 is capable of controlling the average delay to be at a predetermined desired delay, the power saving state it selects is determined when the system gives the desired delay. When the load is high, if the determined state is the deep sleep state, the delay of the data frame may be greatly increased due to a long state transition time.
Nowadays, interactive applications such as e-commerce, web search and transaction systems have higher requirements on the completion time of data transmission, but the huge energy consumption is not negligible. Therefore, a strategy based on 40-100Gbps ethernet design should achieve better power savings while ensuring that the tail delay does not exceed a predetermined delay constraint. However, none of the existing strategies achieve a good compromise in terms of energy saving and control delay.
Disclosure of Invention
The problem that this patent will be solved is: how to provide a 40-100Gbps Ethernet energy-saving strategy based on prediction, and the delay is controlled within a desired value, and meanwhile, the better energy-saving effect is achieved compared with the existing energy-saving strategy.
In order to solve the problems, the invention provides a 40-100Gbps Ethernet energy-saving strategy implementation method based on prediction, which comprises the following steps:
the method comprises the following steps: an initialization phase, in which the number N of data frames transmitted in the first period is directly counted1The time length of the first period is equal to a preset expected period length ETC;
step two: counting the number N of data frames arriving in the last period at the beginning of the ith periodi-1Wherein i is more than or equal to 2;
step three: predicting a number N 'of data frames that will arrive within a currently expected period duration ETC'iAnd calculating transmission N 'according to link bandwidth'iTime tau 'required by data frame'i；
Step four: according to τ'iSelecting the energy-saving state expected to be used in the current period: if τ'iIf the value of (c) is greater than or equal to the Threshold, then the fast wake energy saving state is used in the current cycle by:
a1, passing T after the period startsAtoFWhen the energy-saving Ethernet EEE is in the fast awakening state, the energy-saving Ethernet EEE enters the fast awakening state; wherein, TAtoFRepresents the time required for the EEE to transition from the normal operating state to the fast wake-up state;
a2 at ETC- τ'i-TFtoAAt the moment, EEE is from fastThe fast awakening state is recovered to a normal working state, and after all the cached data frames are transmitted, the next period is entered and the second step is executed; wherein, TFtoARepresenting the time required for the EEE to transition from the fast wake-up state to the normal operating state;
if τ'iIs less than Threshold, the deep sleep power saving state is used in the current cycle by:
b1, T after period startAtoDThe EEE enters a deep sleep state; wherein, TAtoDRepresenting the time required for the EEE to transition from the normal operating state to the deep sleep state;
b2 at ETC- τ'i-TDtoAAt the moment, the EEE is recovered to a normal working state from a deep sleep state, and after all the cached data frames are completely transmitted, the next period is entered and the step two is executed; wherein, TDtoAIndicating the time required for the EEE to transition from the deep sleep state to the normal operating state.
In the method, in the first step, the expected period length ETC is based on the reference expected 99 percent data frame transmission delay T99To set, ETC and T99The following formula is satisfied:
Figure BDA0001918175180000041
wherein λ is a link sending rate, μ is a service rate, and τ is a transmission time of a data frame actually arriving in each period.
In the third step, the number of data frames N 'arriving within the ith expected period duration is predicted'iIs a reaction of Ni-1Is input into an autoregressive moving average model or a predicted value obtained by an exponential smoothing method, Ni-1Is Ti-1Number of data frames, T, of inner transmissioni-1Is the length of the i-1 th cycle.
In the method, the calculation formula of the autoregressive moving average model is as follows:
Figure BDA0001918175180000042
Figure BDA0001918175180000043
Tiis the length of the i-th cycle, NiIs TiNumber of data frames, x, of inner transmissioniIs XiA predicted value of0、b0、ε0Are all 0, in the ith iteration, ai、bi、xiBy adding a toi-1、bi-1、xi-1Is input to a recursive augmented least squares algorithm to obtain the value of (c).
In the method, the calculation formula of the exponential smoothing method is as follows:
Figure BDA0001918175180000044
Figure BDA0001918175180000045
α is a weight, TiIs the length of the i-th cycle, NiIs TiThe number of data frames transmitted in.
In the method, in the fourth step, the Threshold is obtained by solving the following equation about the unknown τ:
Figure BDA0001918175180000046
wherein, the left side of the equation represents the energy consumption in the fast awakening state, and the right side of the equal sign represents the energy consumption in the deep sleep state; t isfIndicating T in fast wake-up stateAtoFAnd TFtoAThe sum of (1); t isdIndicating T in deep sleep stateAtoDAnd TDtoAThe sum of (a) and (b),solving the equation to obtain the Threshold value Threshold as the constant ETC-9.34 mu s.
The energy-saving strategy based on the predicted 40-100Gbps Ethernet can periodically select the optimal energy-saving state for the EEE and control the time when the EEE leaves the energy-saving state on the premise of ensuring that the delay of the data frame tail does not exceed the preset expected delay, and can achieve better energy-saving effect compared with the existing energy-saving strategy. And a better compromise can be obtained in the aspects of energy conservation and tail delay control.
Drawings
FIG. 1 is a diagram of a 40-100Gbps Ethernet power saving policy state transition based on prediction.
FIG. 2 is a state transition diagram of the Dual-Mode policy in 40-100Gbps Ethernet.
FIG. 3 is a state transition diagram of FC policy in a 40-100Gbps Ethernet network.
FIG. 4 is a state transition diagram of the FC-SSHI policy in 40-100Gbps Ethernet.
FIG. 5 is a state transition diagram of the FC-DT policy in a 40-100Gbps Ethernet network.
FIG. 6 is a graph comparing energy consumption, mean delay, tail delay for different EEE strategies at Poisson flow.
FIG. 7 is a graph comparing energy consumption, mean delay, tail delay for different EEE strategies at self-similar flow.
Fig. 8 is a comparison graph of simulation at real flow rate.
Detailed Description
The following describes in further detail embodiments of the present invention with reference to the accompanying drawings.
The 40-100Gbps Ethernet energy-saving strategy based on prediction selects the optimal energy-saving state for EEE and controls the time when the EEE leaves the optimal energy-saving state on the premise of ensuring that the delay of the data frame tail does not exceed the preset expected delay. FIG. 1 is a diagram of a 40-100Gbps Ethernet power saving policy state transition based on prediction. As shown in fig. 1, the ethernet power saving policy based on the prediction operates periodically. The reference procedure for the desired cycle length ETC setting is as follows:
delay profile establishment for intra-period framesAnalyzing the model to obtain Laplace transform T of frame response timer(s)：
Figure BDA0001918175180000061
Wherein, TQ(s) is the laplacian transform of the waiting time of the data frame in the queue, and d(s) is the laplacian transform of the service time.
And deducing a relation between tail delay and cycle length through inverse Laplace transform:
Figure BDA0001918175180000062
the appropriate ETC length is configured with reference to the above formula.
According to the formula
Figure BDA0001918175180000063
Calculating the critical value with the same energy consumption in two energy-saving states: τ ETC-9.34.
When there is no data frame in the link, the current cycle ends and the next cycle is started. At this point, the link will transition from an active state to some power saving state. At the beginning of the ith (i ≧ 2) period, counting the number N of data frames arriving in the last periodi-1The number of data frames N 'to be reached in the ith period is predicted by using ARMA (autoregressive moving average model) or EWMA (exponential smoothing method)'i. The formula for ARMA is as follows:
Figure BDA0001918175180000064
Figure BDA0001918175180000065
Tiis the length of the i-th cycle, NiIs TiNumber of data frames, x, of inner transmissioniIs XiA predicted value of0、b0、ε0Are all 0, in the ith iteration, ai、bi、xiBy adding a toi-1、bi-1、xi-1Is input to a recursive augmented least squares algorithm to obtain the value of (c). When EWMA is used, N 'is present'iETC EWMA (i), wherein
Figure BDA0001918175180000071
Then according to a formula tau'i＝N′i*TtransCalculating of Transmission N'iTime tau 'required by data frame'iWherein, TtransWhich represents the average transmission time of each frame, is determined by the link bandwidth.
If τ'iIf ETC-9.34, EEE passes T at the beginning of the cycleAtoFFrom the active state into the fast wake-up state; the EEE stays ETC-tau 'in the energy-saving state'i-TFtoA-TAtoFAfter time TFtoAThe wake-up time of the data transmission system is converted from a fast wake-up state to a working state, the data frames buffered in the queue are transmitted, the period is finished after the transmission is finished, and the link is converted from the working state to a dormant state. If τ'iETC-9.34 or less, the link passes through T at the beginning of the periodAtoDFrom the active state to the deep sleep state; the EEE stays ETC-tau 'in the energy-saving state'i-TDtoA-TAtoDAfter time TDtoAAnd the wake-up time is converted into a working state from a deep sleep state, the data frames buffered in the queue are transmitted, and the period is finished after the transmission is finished.
In order to further verify the performance of the strategy, the invention realizes and tests the 40-100Gbps Ethernet energy-saving strategy based on prediction on the NS3 platform. For comparison, the Dual-Mode, FC-SSHI and FC-DT strategies were also implemented on the NS3 platform. Experiments show that the prediction-based Ethernet energy-saving strategy can obtain better compromise between energy saving and control delay compared with other strategies.
P-ARMA and P-EWMA in FIG. 6 are the results obtained by the autoregressive moving average model and exponential smoothing method, respectively, of the present invention. It can be seen that the ethernet power saving strategy based on prediction can achieve better power saving effect when the cycle length ETC is 25 μ s, and at the same time, the tail delay can be controlled below 25 μ s, and the tail delay is lower than the FC-DT strategy in the case of high load. While the other three strategies do a relatively poor compromise in terms of energy saving and delay. Fig. 7 shows a comparison graph of the simulation of power saving and delay for several strategies at self-similar flow. The three strategies of Dual-Mode, FC and FC-SSHI are low in delay but low in energy-saving efficiency. While the FC-DT strategy has slightly better energy-saving efficiency than the present invention under the condition of larger arrival rate, the present invention has obvious effect on controlling delay, particularly average delay. The prediction error of the EWMA method may be very large when the arrival rate is large, compared to the ARMA method. Therefore, the difference in delay between P-EWMA and P-ARMA becomes large. However, their average power consumption is almost the same. Overall, P-ARMA performs better than several other strategies. Fig. 8 shows the result of the verification using the real traffic. As shown, although the power saving effect and average delay of the predicted-based Ethernet power saving strategy are not much different from the FC-DT strategy, the tail delay of the predicted-based Ethernet power saving strategy is lower than that of the FC-DT strategy. While other strategies, such as Dual-Mode, FC, and FC-SSHI, consume high power despite their small delays. This is consistent with the previous conclusions. Therefore, the 40-100Gbps ethernet power saving strategy based on prediction is superior to other power saving strategies.

Claims (6)

1. A40-100 Gbps Ethernet energy-saving strategy implementation method based on prediction is characterized by comprising the following steps:
the method comprises the following steps: an initialization phase, in which the number N of data frames transmitted in the first period is directly counted1The time length of the first period is equal to a preset expected period length ETC;
step two: at the beginning of the ith cycle, count the last cycleNumber of data frames up to Ni-1Wherein i is more than or equal to 2;
step three: predicting a number N 'of data frames that will arrive within a currently expected period duration ETC'iAnd calculating transmission N 'according to link bandwidth'iTime tau 'required by data frame'i；
Step four: according to τ'iSelecting the energy-saving state expected to be used in the current period: if τ'iIf the value of (c) is greater than or equal to the Threshold, then the fast wake energy saving state is used in the current cycle by:
a1, passing T after the period startsAtoFWhen the energy-saving Ethernet EEE is in the fast awakening state, the energy-saving Ethernet EEE enters the fast awakening state; wherein, TAtoFRepresents the time required for the EEE to transition from the normal operating state to the fast wake-up state;
a2 at ETC- τ'i-TFtoAAt the moment, the EEE is recovered to a normal working state from the quick awakening state, and after all the cached data frames are completely transmitted, the next period is entered and the step two is executed; wherein, TFtoARepresenting the time required for the EEE to transition from the fast wake-up state to the normal operating state;
if τ'iIs less than Threshold, the deep sleep power saving state is used in the current cycle by:
b1, T after period startAtoDThe EEE enters a deep sleep state; wherein, TAtoDRepresenting the time required for the EEE to transition from the normal operating state to the deep sleep state;
b2 at ETC- τ'i-TDtoAAt the moment, the EEE is recovered to a normal working state from a deep sleep state, and after all the cached data frames are completely transmitted, the next period is entered and the step two is executed; wherein, TDtoAIndicating the time required for the EEE to transition from the deep sleep state to the normal operating state.
2. The method according to claim 1, wherein in step one, the desired period length ETC is based on a desired 99 percent bit data frame transmission delay T99To set, ETC and T99The following formula is satisfied:
Figure FDA0002658138910000021
wherein λ is a link sending rate, μ is a service rate, and τ is a transmission time of a data frame actually arriving in each period.
3. The method of claim 1 wherein in step three, the number of data frames N 'arriving within the ith desired period duration is predicted'iIs a reaction of Ni-1Is input into an autoregressive moving average model or a predicted value obtained by an exponential smoothing method, Ni-1Is Ti-1Number of data frames, T, of inner transmissioni-1Is the length of the i-1 th cycle.
4. The method of claim 3, wherein the autoregressive moving average model is calculated as follows:
Figure FDA0002658138910000022
Figure FDA0002658138910000023
Tiis the length of the i-th cycle, NiIs TiNumber of data frames, x, of inner transmissioniIs XiA predicted value of0、b0、ε0Are all 0, in the ith iteration, ai、bi、xiBy adding a toi-1、bi-1、xi-1Is input to a recursive augmented least squares algorithm to obtain the value of (c).
5. The method of claim 3, wherein the exponential smoothing method is calculated as follows:
Figure FDA0002658138910000024
Figure FDA0002658138910000031
α is a weight, TiIs the length of the i-th cycle, NiIs TiThe number of data frames transmitted in.
6. The method according to claim 1, wherein in the fourth step, the Threshold value Threshold is obtained by solving the following equation for the unknown τ:
Figure FDA0002658138910000032
wherein, the left side of the equation represents the energy consumption in the fast awakening state, and the right side of the equal sign represents the energy consumption in the deep sleep state; t isfIndicating T in fast wake-up stateAtoFAnd TFtoAThe sum of (1); t isdIndicating T in deep sleep stateAtoDAnd TDtoASolving the equation to obtain the Threshold value Threshold as the constant ETC-9.34 mu s.
CN201811582071.1A 2018-12-24 2018-12-24 40-100Gbps Ethernet energy-saving strategy implementation method based on prediction Active CN109617704B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201811582071.1A CN109617704B (en) 2018-12-24 2018-12-24 40-100Gbps Ethernet energy-saving strategy implementation method based on prediction

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201811582071.1A CN109617704B (en) 2018-12-24 2018-12-24 40-100Gbps Ethernet energy-saving strategy implementation method based on prediction

Publications (2)

Publication Number Publication Date
CN109617704A CN109617704A (en) 2019-04-12
CN109617704B true CN109617704B (en) 2021-02-02

Family

ID=66010556

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201811582071.1A Active CN109617704B (en) 2018-12-24 2018-12-24 40-100Gbps Ethernet energy-saving strategy implementation method based on prediction

Country Status (1)

Country Link
CN (1) CN109617704B (en)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112399531B (en) * 2019-08-16 2022-05-06 华为技术有限公司 Equipment awakening method, device and system
CN112887221B (en) * 2021-01-12 2022-11-04 中南大学 Periodic energy-saving method based on generalized predictive control
CN115865543B (en) * 2022-11-30 2024-04-05 中南大学 Periodic energy saving method based on optimal control

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101995823A (en) * 2010-09-28 2011-03-30 吴伪亮 Energy-saving control method based on statistical forecasting technology
CN104285407A (en) * 2012-05-11 2015-01-14 大陆汽车有限责任公司 Method for transmitting data between nodes of motor vehicle by using Ethernet-AVB transport protocol and control unit configured to carry out said method
CN108964931A (en) * 2018-07-12 2018-12-07 长春工业大学 A kind of downstream packet scheduling Ethernet energy-saving scheme based on mixing sleep pattern

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7398408B2 (en) * 2004-11-24 2008-07-08 Conexant Systems, Inc. Systems and methods for waking up wireless LAN devices

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101995823A (en) * 2010-09-28 2011-03-30 吴伪亮 Energy-saving control method based on statistical forecasting technology
CN104285407A (en) * 2012-05-11 2015-01-14 大陆汽车有限责任公司 Method for transmitting data between nodes of motor vehicle by using Ethernet-AVB transport protocol and control unit configured to carry out said method
CN108964931A (en) * 2018-07-12 2018-12-07 长春工业大学 A kind of downstream packet scheduling Ethernet energy-saving scheme based on mixing sleep pattern

Also Published As

Publication number Publication date
CN109617704A (en) 2019-04-12

Similar Documents

Publication Publication Date Title
CN109617704B (en) 40-100Gbps Ethernet energy-saving strategy implementation method based on prediction
CN113612843B (en) MEC task unloading and resource allocation method based on deep reinforcement learning
Wen et al. Energy-optimal mobile application execution: Taming resource-poor mobile devices with cloud clones
WO2023040022A1 (en) Computing and network collaboration-based distributed computation offloading method in random network
CN103889001B (en) A kind of adaptive load balancing method based on future load prediction
CN101272557B (en) Mobile station dormancy mode energy-saving method
US20220417156A1 (en) Network burst load evacuation method for edge servers
CN103108039B (en) Service quality guarantee method in low-energy cluster environment
Herrería-Alonso et al. How efficient is energy-efficient ethernet?
CN113220364A (en) Task unloading method based on vehicle networking mobile edge computing system model
CN106941712A (en) A kind of micro-base station power-economizing method cooperated based on traffic forecast and macro base station
CN115066006A (en) Base station dormancy method, equipment and medium based on reinforcement learning
CN111884947B (en) Data packet management method based on information age at receiving end
CN104754063A (en) Local cloud computing resource scheduling method
CN101217484A (en) Resource distributing method and system
Rodríguez-Pérez et al. Optimum traffic allocation in bundled energy-efficient Ethernet links
CN105007225A (en) Network dynamic topology control system and method thereof
Penttinen et al. Energy-aware dispatching in parallel queues with on-off energy consumption
CN108075915A (en) A kind of RDMA communication connection pond management methods based on ADAPTIVE CONTROL
CN115865543B (en) Periodic energy saving method based on optimal control
CN114706631A (en) Unloading decision method and system in mobile edge calculation based on deep Q learning
CN110320813A (en) Intelligent electrical appliance control and device based on Internet of Things operating system
CN107729070B (en) Virtual machine scheduling system and method based on double speed and work dormancy
CN114040477A (en) Base station energy saving method, device, equipment and product
CN103916261B (en) Determine the method and control device for entering the quantity of network equipment of energy-saving mode

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant