CN108512708B - Cache calculation method and device - Google Patents
Cache calculation method and device Download PDFInfo
- Publication number
- CN108512708B CN108512708B CN201710103163.6A CN201710103163A CN108512708B CN 108512708 B CN108512708 B CN 108512708B CN 201710103163 A CN201710103163 A CN 201710103163A CN 108512708 B CN108512708 B CN 108512708B
- Authority
- CN
- China
- Prior art keywords
- link
- time
- delay
- message
- interleaving
- 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
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/08—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
- H04L43/0852—Delays
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L49/00—Packet switching elements
- H04L49/90—Buffering arrangements
- H04L49/9047—Buffering arrangements including multiple buffers, e.g. buffer pools
- H04L49/9052—Buffering arrangements including multiple buffers, e.g. buffer pools with buffers of different sizes
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Environmental & Geological Engineering (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention provides a method and a device for cache calculation, wherein the method is suitable for multilinks which comprise at least one unstable link and at least one stable link and comprises the following steps: acquiring an average message time interval of a multilink, a first link time delay of an unstable link and a second link time delay of a stable link; the method comprises the steps of calculating the number of message caches according to an average message time interval, a first link time delay and a second link time delay, reasonably calculating the number of the message caches according to different time delays, solving the problem of disorder caused by the time delays of different links, reasonably utilizing resources and expenses, finishing order preservation in time when packet loss occurs according to the calculated number of the message caches, triggering rapid retransmission of TCP, preventing the TCP from overtime caused by the order preservation, ensuring the transmission rate and optimizing the multilink transmission performance.
Description
Technical Field
The present invention relates to the field of communications, and in particular, to a method and an apparatus for cache calculation.
Background
The problem of disorder is a relatively common phenomenon in a network load balancing scenario, and for different link devices or uplink modes, due to inconsistent transmission delays in the network, a situation may occur in which a message sent first arrives at a terminal device instead of the message sent last. For example, when a DSL (Digital Subscriber Line) and an LTE dual link work simultaneously, because the delay of the DSL link is not consistent with that of the LTE, it may happen that data transmitted first by the LTE (Long Term Evolution) link arrives later than a data packet transmitted by the DSL link.
A common order preserving method generally caches a message, and sends out the cached message when a to-be-timed message expires or a positive-order message arrives, for example, a network device may start a timer or reserve a certain cache, cache a DSL-arrived message first, and send out the ordered message when a LTE message arrives or the timer expires.
The method is suitable for the situation that the time delay of two links is stable, such as DSL, ETH (Ethernet) link and the like. However, as described above, when the DSL and the LTE link operate simultaneously, due to the instability of the delay of the LTE link, the message caching method cannot determine the maximum cache size, and if the cache is too small, the order preservation purpose cannot be achieved, and if the cache is too large, the system load and overhead of the network device are increased, which causes unnecessary waste. Of course, if the method of using the timer is used, the expiration time of the timer cannot be preset, and the problem that the number of the buffers cannot be preset still exists.
Disclosure of Invention
In view of this, embodiments of the present invention provide a method and an apparatus for cache calculation, which solve the problem that the cache size of the order-preserving queue cannot be correctly estimated due to dynamic change of time delay of a link.
In order to solve the above problem, in a first aspect, an embodiment of the present invention provides a method for cache calculation, which is applicable to multiple links, where the multiple links include at least one unstable link and at least one stable link, and the method includes:
acquiring the average message time interval of a multilink, the first link time delay of an unstable link and the second link time delay of a stable link;
and calculating the number of message caches according to the average message time interval, the first link time delay and the second link time delay.
Optionally, the obtaining the average packet time interval of the multilink includes:
acquiring the number of messages passing through a multilink within a first preset time by a sampling method;
calculating the average message time interval, wherein the calculation formula is as follows:
Δt＝ΔT/m
wherein, Δ T is an average message time interval, Δ T is a first preset time, and m is the number of messages, wherein the first preset time is less than the longest link delay, and the number of messages is less than the number of message caches.
Optionally, the calculation formula of the second link delay is:
ΔT D ＝(ΔT time of repacking -ΔT Uplink interleaving -ΔT D downlink interleaving )/2+ΔT D downlink interleaving
Wherein, Δ T D For the second link delay, Δ T Time of return package For the time of repacking, Δ T Uplink interleaving For uplink interleaving time, Δ T D downlink interleaving Is the downlink interleaving time.
Optionally, the method for calculating the first link delay includes:
after receiving a group of messages sent by a sending end, sending a confirmation message to the sending end by a stable link, and recording the sending time of the stable link as a first time, wherein the confirmation message comprises an expected Sequence Number;
recording a second time when a next group of messages which are sent by the sending end through the unstable link and contain expected serial numbers are received;
and calculating the data flow delay of the unstable link according to the second time, the first time and the uplink delay of the stable link, wherein the calculation formula is as follows:
ΔT L1 ＝T2-T1-ΔT D2
wherein, delta T L1 For data stream delay, T1 is a first time, T2 is a second time, Δ T D2 Uplink delay for a stable link;
calculating the average value of the data stream time delay according to the acquired data stream time delay of at least two groups of unstable links, wherein the average value is the first link time delay delta T L 。
Optionally, the formula for calculating the number of the packet caches is as follows:
n≥(ΔT L -ΔT D )/Δt
wherein n is the number of message caches, and delta T L For the first link delay, Δ T D For the second link delay, Δ t is the average message time interval.
Optionally, the method further comprises:
and if the first link time delay of the unstable link meets the updating rule, recalculating the message cache number.
Optionally, if the first link delay of the unstable link satisfies the update rule, recalculating the number of packet caches includes:
and if the latest acquired time delay and the last acquired time delay in the acquired time delays of the plurality of first links meet the updating rule, recalculating the number of the message caches.
Optionally, the update rule is:
T n ≥T n-1 * Beta or T n ≤T n-1 *β
Wherein, T n For the latest acquired first link delay, T n-1 For the first link delay obtained last time, β is an influence factor, and its value is determined by the first link delay of the unstable link.
Optionally, the method further comprises:
and if the number of the obtained reports Wen Huancun in the current order-preserving queue exceeds the calculated number of the message caches, ending the order-preserving process.
In a second aspect, the present invention further provides an apparatus for cache computing, including:
an acquisition module: the method comprises the steps of obtaining an average message time interval of a multilink, a first link time delay of an unstable link and a second link time delay of a stable link;
a calculation module: and the method is used for calculating the number of message caches according to the first link time delay, the second link time delay and the average message time interval.
Optionally, the obtaining module includes:
a counting unit: the method comprises the steps of obtaining the number of messages passing through a multilink within a first preset time by a sampling method;
a first processing unit: and the average message time interval is calculated according to the first preset time and the number of the messages.
Optionally, the obtaining module further includes a second processing unit, configured to calculate a second link delay, where the calculation formula is:
ΔT D ＝(ΔT time of repacking -ΔT Uplink interleaving -ΔT D downstream interleaving )/2+ΔT D downstream interleaving
Wherein, delta T Time of return package For the time of repacking, Δ T Uplink interleaving For uplink interleaving time, Δ T D downlink interleaving Is the downlink interleaving time.
Optionally, the obtaining module further includes a third processing unit, configured to calculate the first link delay, where the third processing unit includes:
a first recording subunit: the system comprises a stable link, a sending end and a receiving end, wherein the stable link is used for sending a group of messages to the sending end through the stable link after receiving the group of messages sent by the sending end, and recording the sending time of the stable link as first time, wherein the confirmation messages contain expected serial numbers;
a second recording subunit: the second time is recorded after the next group of messages which are sent by the sending end through the unstable link and contain expected serial numbers is received;
a first calculation subunit: and calculating the data flow delay of the unstable link according to the second time, the first time and the uplink delay of the stable link, wherein the calculation formula is as follows:
ΔT L1 ＝T2-T1-ΔT D2
wherein, delta T L1 For data stream delay, T1 is a first time, T2 is a second time, Δ T D2 Uplink delay for a stable link;
a second calculation subunit: calculating the average value of the data stream time delay according to the acquired data stream time delay of at least two groups of unstable links, wherein the average value is the first link time delay delta T L 。
Optionally, the apparatus further comprises:
an updating module: and the method is used for recalculating the message cache number if the first link time delay of the unstable link meets the updating rule.
Optionally, the update module includes:
a recalculation unit: and if the latest acquired time delay and the last acquired time delay in the acquired time delays of the plurality of first links meet the updating rule, recalculating the number of the message caches.
Optionally, the apparatus further comprises:
a termination module: and if the number of the obtained reports Wen Huancun in the current order-preserving queue exceeds the calculated number of the message caches, ending the order-preserving process.
In summary, the present invention calculates the number of buffered packets through the average packet time interval of the multilink, the first link delay of the unstable link, and the second link delay of the stable link, according to the average packet time interval, the first link delay, and the second link delay, and can reasonably calculate the number of buffered packets according to different delays when the multilink includes the unstable link, thereby reasonably utilizing resources and overhead, solving the problem of disorder caused by delays of different links, and meanwhile, according to the calculated number of buffered packets, when a packet is lost, ending the order preservation in time, triggering fast retransmission of TCP (Transmission Control Protocol), preventing TCP timeout due to the timeout of the order preservation, ensuring Transmission rate, and optimizing the Transmission performance of the multilink.
Drawings
FIG. 1 is a flow chart of a method of cache calculation according to an embodiment of the present invention;
FIG. 2 is another flow chart of a method of cache calculation according to an embodiment of the present invention;
FIG. 3 is another flow chart of a method of cache calculation according to an embodiment of the present invention;
FIG. 4 is another flow chart of a method of cache calculation according to an embodiment of the present invention;
fig. 5 is a schematic structural diagram of a cache computing apparatus according to an embodiment of the present invention.
Detailed Description
In order to make the technical problems, technical solutions and advantages of the present invention more apparent, the following detailed description is given with reference to the accompanying drawings and specific embodiments.
Referring to fig. 1, a flow chart of a method of cache computation is shown, the method being applicable to a multilink, wherein the multilink includes at least one unstable link and at least one stable link, and comprising the steps of:
s101, obtaining an average message time interval of a multilink, a first link time delay of an unstable link and a second link time delay of a stable link.
In this embodiment, a calculation method is described by taking simultaneous operation of DSL and LTE as an example, and of course, the method can also be used in a scenario of multi-link forwarding, where at least one unstable link is in the link, for example, LTE in this embodiment is unstable, that is, the delay of the link may change.
The message is a data unit exchanged and transmitted in the network, that is, a data block to be sent by the station at one time, and the message includes complete data information to be sent; the link refers to a link which is a physical line from one node to an adjacent node without any other switching node in the middle.
The delay refers to the time required for a message or packet to travel from one end of a network to another.
In this embodiment, the LTE link (unstable link) delay is defined as a first link delay Δ T L Defining the DSL link (stable link) delay as the second link delay Δ T D 。
In this embodiment, both the first link delay and the second link delay are round trip delays, where Δ T is the delay of the first link (stable link) D Measurement is performed by using an ICMP (Internet Control Message Protocol).
For example, the measurement can be performed by a ping tool, a sending end of the ping tool sends an ICMP request message to a designated receiving end, the receiving end returns a response message immediately after receiving the message, and the sending end calculates the round-trip network transmission delay of the message between the two ends by sending and receiving clocks at the two ends of the message after receiving the response message.
Specifically, in the present embodiment, the time delay Δ T of the stable link (for example, DSL link) D The packet return time and the uplink and downlink interleaving time of the DSL can be obtained at the convergence point of the double links, and the convergence point refers to the convergence point of the stable link and the unstable link.
ΔT Time of return package ＝ΔT D uplink +ΔT Uplink interleaving +ΔT D goes downward +ΔT D downlink interleaving
ΔT D1 ＝ΔT D goes downward +ΔT D downlink interleaving
ΔT D2 ＝ΔT D uplink +ΔT D uplink interleaving
Wherein, Δ T Time of return package For the back-packet time, the round-trip delay of ICMP can be detected, Δ T D1 For downlink time delay, Δ T D2 For uplink time delay, uplink refers to the number of user side transmitting to network sideAccordingly, downlink refers to the network side sending data to the user side, Δ T Uplink interleaving And Δ T D downstream interleaving Respectively uplink and downlink interleaving times, which can be obtained in the negotiation context, Δ T D uplink And Δ T D goes downward The uplink transmission time and the downlink transmission time of the data packet are respectively.
Neglecting the time consumed by the system, Δ T D uplink ＝ΔT D goes downward In this embodiment, the second link delay is a downlink delay, that is, Δ T D ＝ΔT D1 And finally, the calculation formula of the time delay of the second link is as follows:
ΔT D ＝(ΔT time of repacking -ΔT Uplink interleaving -ΔT D downlink interleaving )/2+ΔT D downlink interleaving
In this embodiment, the dual link is encapsulated in a GRE (Generic Routing Encapsulation) tunnel, and the delay of the unstable link (DSL link) can be measured by a Sequence Number (Sequence Number) of the GRE and an Acknowledgement (ACK).
In this embodiment, after receiving a group of messages sent by a sending end, the stable link sends an acknowledgement message to the sending end, and the sending time of the stable link is recorded as a first time T1.
Specifically, a sending window may be made when the sending end sends data to the unstable link, the size of the sending window defaults to 10 messages, of course, the size of the window may also be other numbers of messages, and the sending end is located at an intersection of the stable link and the unstable link, and may be a bearer network device.
After the sending end finishes sending a group of 10 messages (for example, messages 1 to 10), the sending window is exhausted, no message is sent any more, and the subsequent messages are cached.
After receiving the message, the receiving end device sends an acknowledgement message to the sending end through the stable link, for example, by applying for an ACK carrying an expected sequence number, for example, seqnum =11, the sending end sends the message including the ACK from the stable link (DSL link) to the sending end, and after receiving the ACK message, the sending end re-initializes the sending window, and sends a next group of messages including the expected sequence number through the unstable link, for example, sends 11 th to 20 th messages.
And after receiving the next group of group messages containing the expected serial number and sent by the sending end, the receiving end equipment records a second time T2.
And calculating the data flow delay of the unstable link according to the second time, the first time and the uplink delay of the stable link, wherein the calculation formula is as follows:
ΔT L1 ＝T2-T1-ΔT D2
wherein, delta T L1 For data stream delay, T1 is a first time, T2 is a second time, Δ T D2 In order to stabilize the uplink delay of the link, the uplink delay may be the uplink delay of the stable link obtained based on the ICMP, or may be obtained in other manners.
Obtaining data stream time delay of at least two groups of unstable links, and calculating the average value of the data stream time delay, wherein the average value is the first link time delay delta T L . For example, 10 data streams may be transmitted, and the 2 nd to 10 th data stream time delays, Δ T, may be calculated sequentially after the 10 th data streams are transmitted L1 …ΔT L10 That is, the sampling value of the first link time delay is averaged to obtain the first link time delay delta T L ：
ΔT L ＝(ΔT L1 +…+ΔT L10 )/10
In this embodiment, the receiving end device may be an access terminal CPE (Customer Premise Equipment).
In this embodiment, one packet is forwarded from the LTE at time T1, and a plurality of packets are forwarded from the DSL link at time n, n +1, so as to ensure that the serial number of the packet is not disordered, which is equivalent to the requirement:
ΔT L +T1≤ΔT D +T n+1
however, if T in the above formula n+1 By calculating the time interval of all messages, the time interval Δ t of each message needs to be known i Therefore, it is difficult to calculate the number n of buffers, and therefore, in this embodiment, the number of messages passing through within a certain time is estimated by using the multilink average message time interval.
And S102, calculating the number of message caches according to the average message time interval, the first link time delay and the second link time delay.
In this embodiment, the average packet time interval is substituted into the calculation formula of Tn to obtain:
T n+1 ＝T1+Δt 1 +Δt 2 +……+Δt n +Δt n+1 ＝T1+n*Δt
namely:
ΔT L +T1≤ΔT D +T1+n*Δt
therefore, the calculation formula of the number of the message caches is as follows:
n≥(ΔT L -ΔT D )/Δt
wherein n is the number of message caches, and delta T L For the first link delay, Δ T D And delta t is the time delay of the second link, and is the average message time interval.
In summary, in this embodiment, by obtaining the average packet time interval, the first link delay, and the second link delay, and calculating the number of buffers according to the average packet time interval and the link delay difference, the number of buffers of a packet can be reasonably calculated according to different delays, thereby solving the problem of disorder caused by delays of different links, and reasonably utilizing resources and overhead.
Second embodiment
Referring to fig. 2, another flow chart of a method of cache computation is shown, comprising the steps of:
s201, obtaining an average message time interval of a multilink, a first link time delay of an unstable link and a second link time delay of a stable link.
In this embodiment, the method is described by taking the DSL and LTE dual links as an example, and of course, the method can also be used in a scenario of multi-link forwarding, where at least one of the links is unstable, for example, LTE in this embodiment is unstable, that is, the delay of the link may change.
The messages are data units exchanged and transmitted in the network, and the messages contain complete data information to be sent.
Defining the LTE link (unstable link) delay as the first link delay Delta T L Defining the DSL link (stable link) delay as the second link delay Δ T D 。
In this embodiment, both the first link delay and the second link delay are round-trip delays, where the first link (stable link) delay is measured by ICMP, for example, network transmission delay of a message round-trip between two ends may be measured by a ping tool.
Specifically, in the present embodiment, the time delay Δ T of the stable link (for example, DSL link) D The packet return time and the uplink and downlink interleaving time of the DSL can be calculated at the convergence point of the double links, wherein the convergence point is the convergence point of the stable link and the unstable link.
ΔT Time of repacking ＝ΔT D uplink +ΔT Uplink interleaving +ΔT D goes downward +ΔT D downlink interleaving
ΔT D1 ＝ΔT D goes downward +ΔT D downlink interleaving
ΔT D2 ＝ΔT D uplink +ΔT D uplink interleaving
Wherein, delta T Time of repacking For the back-packet time, the round-trip delay of ICMP can be detected, Δ T D1 For downlink time delay, Δ T D2 For uplink time delay, uplink refers to the user side sending data to the network side, downlink refers to the network side sending data to the user side, and Δ T Uplink interleaving And Δ T D downstream interleaving Respectively uplink and downlink interleaving times, which can be obtained in the negotiation context, Δ T D uplink And Δ T D goes downward Respectively, the uplink transmission time and the downlink transmission time of the data packet.
Neglecting the time consumed by the system, Δ T D uplink ＝ΔT D goes downward In this embodiment, the second link delay is a downlink delay, that is, Δ T D ＝ΔT D1 And finally, the calculation formula of the time delay of the second link is as follows:
ΔT D ＝(ΔT time of return package -ΔT Uplink interleaving -ΔT D downlink interleaving )/2+ΔT D downlink interleaving
In this embodiment, the dual link is encapsulated in a GRE tunnel, and the delay of the unstable link (DSL link) can be measured through seqnum of the GRE and the reply ACK.
In this embodiment, after receiving a group of messages sent by a sending end, the stable link sends an acknowledgement message to the sending end, and the sending time of the stable link is recorded as a first time T1.
Specifically, a sending window may be made when the sending end sends data to the unstable link, the size of the sending window defaults to 10 messages, of course, the size of the window may also be other numbers of messages, and the sending end is located at an intersection of the stable link and the unstable link, and may be a bearer network device.
After the sending end sends a group of 10 messages, the sending window is exhausted, the messages are not sent any more, and the subsequent messages are cached.
After receiving the message, the receiving end device sends an acknowledgement message to the sending end through the stable link, for example, by applying for an ACK, which carries an expected serial number, the sending end sends the message containing the ACK from the stable link (DSL link) to the sending end, and after receiving the ACK message, the sending end re-initializes the sending window, and sends the next group of messages containing the expected serial number through the unstable link.
And after receiving the next group of group messages containing the expected serial number and sent by the sending end, the receiving end equipment records a second time T2.
And calculating the data flow delay of the unstable link according to the second time, the first time and the uplink delay of the stable link, wherein the calculation formula is as follows:
ΔT L1 ＝T2-T1-ΔT D2
wherein, delta T L1 For data stream delay, T1 is a first time, T2 is a second time, and Δ T D2 In order to stabilize the uplink delay of the link, the uplink delay may be the uplink delay of the stable link obtained based on the ICMP, or may be obtained in other manners.
In this embodiment, the data stream delays of at least two groups of unstable links are obtained, and an average value of the data stream delays is calculated, where the average value is the first link delay Δ T L . For example, 10 data streams may be transmitted, and subsequently calculated in turn to be2-10 data stream delay, Δ T L1 …ΔT L10 The sampling value of the first link time delay is obtained, and the average value is obtained to obtain the time delay of the first link:
ΔT L ＝(ΔT L1 +…+ΔT L10 )/10
in this embodiment, the receiving end device may be an access terminal CPE.
In this embodiment, obtaining the average packet time interval includes the following steps:
and S2011, acquiring the number of messages passing through the multilink within the first preset time by using a sampling method.
In this embodiment, the number of the messages passing through within a certain time (for example, a first preset time) is estimated by using a sampling method, where of course, the first preset time is less than the time delay of the longest link, and the number of the messages passing through within the first preset time is less than the number of the message caches.
S2012, calculating the average message time interval.
In this embodiment, the calculation formula of the average packet time interval is as follows:
Δt＝ΔT/m
wherein, Δ T is the average message time interval, Δ T is the first preset time, and m is the number of messages.
S202, calculating the number of message caches according to the average message time interval, the first link time delay and the second link time delay.
In this embodiment, the calculation formula of the number of message caches is as follows:
n≥(ΔT L -ΔT D )/Δt
wherein n is the number of message caches, and delta T L For the first link delay, Δ T D For the second link delay, Δ t is the average message time interval.
And S203, if the first link delay of the unstable link meets the updating rule, recalculating the message cache number.
In this embodiment, the second link delay Δ T D Is generally relatively fixed, but the first link delay deltat L Is likely to vary with Δ T L Delay increase, number of message buffers nThe minimum value will also increase, causing n to change dynamically with time delay and message time interval, so the value of the number of message caches needs to be updated.
In this embodiment, whether to update the number of the message caches is determined according to the first link delay, and when the first link delay meets the update rule, the number of the message caches is recalculated, so that the number of the cache messages can be dynamically calculated.
In this embodiment, the LTE link is one in which the delay is prone to jitter and the delay is large, so that the rule is determined to be updated according to the delay of the LTE transmission packet.
In summary, in this embodiment, the average packet time interval is obtained by obtaining the number of packets that pass through within the first preset time through the sampling method, and the number of packet caches is calculated according to the obtained first link delay, the obtained second link delay, and the average packet time interval, so that the number of packet caches can be reasonably calculated according to different delays, and meanwhile, when the first link delay meets the update rule, the number of packet caches is recalculated, so that the number of packet caches can be dynamically calculated, the problem of disorder caused by delays of different links is solved, and resources and overhead are reasonably utilized.
Third embodiment
Referring to fig. 3, another flowchart of the cache calculation method according to the embodiment is shown, where the method includes:
s301, obtaining the average message time interval of the multilink, the first link time delay of the unstable link and the second link time delay of the stable link.
In this embodiment, the method is described by taking the DSL and LTE dual links as an example, and of course, the method can also be used in a scenario of multi-link forwarding, where at least one of the links is unstable, for example, LTE in this embodiment is unstable, that is, the delay of the link may change.
The messages are data units exchanged and transmitted in the network, and the messages contain complete data information to be sent.
Defining the LTE link (unstable link) delay as the first link delay Delta T L Defining the DSL link (stable link) delay as the second link delay Δ T D 。
In this embodiment, both the first link delay and the second link delay are round-trip delays, where the first link (stable link) delay is measured by ICMP, for example, network transmission delay of a message round-trip between two ends may be measured by a ping tool.
Specifically, the time delay Δ T of the stable link (e.g., DSL link) in the present embodiment D The packet return time and the uplink and downlink interleaving time of the DSL can be calculated at the convergence point of the double links, wherein the convergence point is the convergence point of the stable link and the unstable link.
ΔT Time of return package ＝ΔT D uplink +ΔT Uplink interleaving +ΔT D goes downward +ΔT D downstream interleaving
ΔT D1 ＝ΔT D goes downward +ΔT D downlink interleaving
ΔT D2 ＝ΔT D uplink +ΔT D uplink interleaving
Wherein, delta T Time of repacking For the back-packet time, the round-trip delay of ICMP can be detected, Δ T D1 For downlink time delay, Δ T D2 For uplink time delay, uplink refers to the user side sending data to the network side, downlink refers to the network side sending data to the user side, and Δ T Uplink interleaving And Δ T D downlink interleaving Respectively uplink and downlink interleaving times, which can be obtained in the negotiation context, Δ T D uplink And Δ T D goes downward Respectively, the uplink transmission time and the downlink transmission time of the data packet.
Neglecting the time consumed by the system, Δ T D uplink ＝ΔT D goes downward In this embodiment, the second link delay is a downlink delay, that is, Δ T D ＝ΔT D1 And finally, the calculation formula of the time delay of the second link is as follows:
ΔT D ＝(ΔT time of return package -ΔT Uplink interleaving -ΔT D downlink interleaving )/2+ΔT D downlink interleaving
In this embodiment, the dual link is encapsulated in a GRE tunnel, and the delay of the unstable link (DSL link) can be measured through seqnum of the GRE and the reply ACK.
In this embodiment, after receiving a group of messages sent by a sending end, the stable link sends an acknowledgement message to the sending end, and the sending time of the stable link is recorded as a first time T1.
Specifically, a sending window is made when a sending end sends data to an unstable link, the size of the sending window defaults to 10 messages, certainly, the size of the window can be other numbers of messages, and the sending end is located at an intersection point of the stable link and the unstable link and can be a carrying network device.
After the sending end sends a group of 10 messages, the sending window is exhausted, the messages are not sent any more, and the subsequent messages are cached.
After receiving end equipment receives a message sent by a sending end, a stable link sends an acknowledgement message to the sending end, for example, by applying for an ACK carrying an expected serial number, the message containing the ACK is sent from the stable link (DSL link) to the sending end, after the sending end receives the ACK message, a sending window is reinitialized, and a next group of messages containing the expected serial number is sent by an unstable link.
And after receiving end equipment receives the next group of messages which are sent by the sending end and contain the expected serial numbers, recording the second time T2.
And calculating the data flow delay of the unstable link according to the second time, the first time and the uplink delay of the stable link, wherein the calculation formula is as follows:
ΔT L1 ＝T2-T1-ΔT D2
wherein, delta T L1 For data stream delay, T1 is a first time, T2 is a second time, Δ T D2 In order to stabilize the uplink delay of the link, the uplink delay may be the uplink delay of the stable link obtained based on the ICMP, or may be obtained in other manners.
In this embodiment, the data stream delays of at least two groups of unstable links are obtained, and an average value of the data stream delays is calculated, where the average value is the first link delay Δ T L . For example, 10 data streams may be transmitted, and the 2 nd to 10 th data stream time delay, Δ T, may be calculated sequentially L1 …ΔT L10 Is thatAnd averaging the sampling values of the first link time delay to obtain the time delay of the first link:
ΔT L ＝(ΔT L1 +…+ΔT L10 )/10
in this embodiment, the receiving end device may be an access terminal CPE.
In this embodiment, the average packet time interval is obtained by a sampling method, that is:
Δt＝ΔT/m
wherein, Δ T is the average message time interval, Δ T is the first preset time, and m is the number of messages.
S302, calculating the number of message caches according to the average message time interval, the first link time delay and the second link time delay.
In this embodiment, the calculation formula of the number of message caches is as follows:
n≥(ΔT L -ΔT D )/Δt
wherein n is the number of message caches, and delta T L For the first link delay, Δ T D For the second link delay, Δ t is the average message time interval.
S303, judging that the latest acquired time delay and the last acquired time delay in the acquired time delays of the plurality of first links meet an updating rule, and if so, entering the step S304; otherwise, the flow ends.
In this embodiment, the first link delay refers to a delay of a link with a delay that is prone to jitter or a delay that is large, for example, the delay of an LTE link, the delay of each packet transmitted from the LTE link is measured and recorded, and the latest recorded delay time T is compared among a plurality of obtained delays n Time delay T from last recording n-1 Wherein, the time recorded last time is taken as a reference.
The update rule is as follows:
T n ≥T n-1 * Beta or T n ≤T n-1 *β
Wherein, T n For the latest recorded time delay, T n-1 Beta is the influence factor for the time delay time of the last recording.
In this embodiment, the latest recorded delay time T n Time delay T from last recording n-1 Compared with the prior art, the time delay time can be prolonged or shortened, if the time delay time recorded for the latest time is greater than the time delay time recorded for the last time, the time delay is prolonged, and the updating rule is as follows:
T n ≥T n-1 *β
if the latest recorded delay time is less than the last recorded delay time, i.e. the delay time is shortened, the update rule is as follows:
T n ≤T n-1 *β
the influence factor β is related to the time delay, and the network with the time delay greater than 1s has no great practical significance, so the time upper limit of the time delay is limited to 1s, and the specific relationship is as follows:
| influencing factor beta | Time interval |
| 500％ | 1ms～10ms |
| 100％ | 11ms～200ms |
| 50％ | 201ms～500ms |
| 20％ | 501ms～1s |
| 10％ | >1s |
S304, recalculating the number of the message caches.
In this embodiment, when the latest acquired time and the previous acquired time satisfy the update rule, the size of the dynamic cache is updated.
In summary, in the embodiment, the number of the message caches is reasonably calculated according to different time delays, and is updated according to the time delay of the unstable link, so that the number of the message caches can be dynamically calculated, resources and overhead are reasonably utilized, and the problem of disorder caused by the time delays of different links is solved.
Fourth embodiment
Referring to fig. 4, another flowchart of the cache calculation method according to the embodiment is shown, where the method includes:
s401, obtaining the average message time interval of the multilink, the first link time delay of the unstable link and the second link time delay of the stable link.
In this embodiment, the method is described by taking the DSL and LTE dual links as an example, and of course, the method can also be used in a scenario of multi-link forwarding, where at least one of the links is unstable, for example, LTE in this embodiment is unstable, that is, the delay of the link may change.
The message is a data unit exchanged and transmitted in the network, and the message contains complete data information to be sent; the link means that the link is a segment of a physical line without any other switching node in between.
Defining the LTE link (unstable link) time delay as a first link time delay delta T L Defining the DSL link (stable link) delay as the second link delay Δ T D 。
In this embodiment, both the first link delay and the second link delay are round-trip delays, where the first link (stable link) delay is measured by ICMP, for example, network transmission delay of a message round-trip between two ends may be measured by a ping tool.
Specifically, the time delay Δ T of the stable link (e.g., DSL link) in the present embodiment D The packet return time and the uplink and downlink interleaving time of the DSL can be calculated at the convergence point of the double links, wherein the convergence point is the convergence point of the stable link and the unstable link.
ΔT Time of return package ＝ΔT D uplink +ΔT Uplink interleaving +ΔT D goes downward +ΔT D downstream interleaving
ΔT D1 ＝ΔT D goes downward +ΔT D downlink interleaving
ΔT D2 ＝ΔT D uplink +ΔT D uplink interleaving
Wherein, delta T Time of return package For the back-packet time, the round-trip delay of ICMP can be detected, Δ T D1 For downlink time delay, Δ T D2 For uplink time delay, uplink refers to the user side sending data to the network side, downlink refers to the network side sending data to the user side, and Δ T Uplink interleaving And Δ T D downlink interleaving Respectively uplink and downlink interleaving times, which can be obtained in the negotiation context, Δ T D uplink And Δ T D goes downward Respectively, the uplink transmission time and the downlink transmission time of the data packet.
Neglecting the time consumed by the system, Δ T D uplink ＝ΔT D goes downward In this embodiment, the second link delay is a downlink delay, that is, Δ T D ＝ΔT D1 And finally, the calculation formula of the time delay of the second link is as follows:
ΔT D ＝(ΔT time of return package -ΔT Uplink interleaving -ΔT D downlink interleaving )/2+ΔT D downlink interleaving
In this embodiment, the dual link is encapsulated in a GRE tunnel, and the delay of the unstable link (DSL link) can be measured through seqnum of the GRE and the reply ACK.
In this embodiment, after receiving a group of messages sent by a sending end, the stable link sends an acknowledgement message to the sending end, and the sending time of the stable link is recorded as a first time T1.
Specifically, a sending window is made when a sending end sends data to an unstable link, the size of the sending window defaults to 10 messages, certainly, the size of the window can be other numbers of messages, and the sending end is located at an intersection point of the stable link and the unstable link and can be a carrying network device.
After the sending end sends a group of 10 messages, the sending window is exhausted, the messages are not sent any more, and the subsequent messages are cached.
After receiving end equipment receives a message sent by a sending end, a stable link sends an acknowledgement message to the sending end, for example, by applying for an ACK carrying an expected serial number, the message containing the ACK is sent from the stable link (DSL link) to the sending end, after the sending end receives the ACK message, a sending window is reinitialized, and a next group of messages containing the expected serial number is sent by an unstable link.
And after receiving the next group of group messages containing the expected serial number and sent by the sending end, the receiving end equipment records a second time T2.
And calculating the data flow delay of the unstable link according to the second time, the first time and the uplink delay of the stable link, wherein the calculation formula is as follows:
ΔT L1 ＝T2-T1-ΔT D2
wherein, Δ T L1 For data stream delay, T1 is a first time, T2 is a second time, and Δ T D2 In order to stabilize the uplink delay of the link, the uplink delay may be the uplink delay of the stable link obtained based on the ICMP, or may be obtained in other manners.
In this embodiment, the data stream delays of at least two groups of unstable links are obtained, and an average value of the data stream delays is calculated, where the average value is the first link delay Δ T L . For example, 10 data streams may be transmitted, and the 2 nd to 10 th data stream time delays, Δ T, may be calculated sequentially after the 10 th data streams are transmitted L1 …ΔT L10 The sampling value of the first link time delay is obtained, and the average value is obtained to obtain the time delay of the first link:
ΔT L ＝(ΔT L1 +…+ΔT L10 )/10
in this embodiment, the receiving end device may be an access terminal CPE.
In this embodiment, the average packet time interval is obtained by a sampling method, that is:
Δt＝ΔT/m
wherein, Δ T is the average message time interval, Δ T is the first preset time, and m is the number of the sampling messages.
S402, calculating the number of message caches according to the average message time interval, the first link time delay and the second link time delay.
In this embodiment, the calculation formula of the number of message caches is as follows:
n≥(ΔT L -ΔT D )/Δt
wherein n is the number of message caches, and delta T L For the first link delay, Δ T D For the second link delay, Δ t is the average message time interval.
S403, judging whether the number of the message caches in the current order preserving queue is larger than the calculated number of the message caches, if so, entering the step S404; otherwise, the flow ends.
The step is used for judging whether packet loss occurs or not, and the order preserving process can be ended when the packet loss occurs. The conventional order preserving method cannot judge when to end the order preserving process, and can only wait until the timer expires or the order preserving queue is full of cache, so that packet loss can be pre-judged in the embodiment.
It should be noted that the calculated number of the message caches may refer to the number of the message caches in step S402, or may be the number of the message caches recalculated after the update rule is satisfied, and if the calculated number of the message caches is recalculated, the steps S402 and S403 further include the step of recalculating the number of the message caches if the delay of the first link of the unstable link satisfies the update rule.
And S404, ending the order preserving process.
In this embodiment, when the number of message buffers exceeds the calculated number of buffers, the order preservation is finished in time, and the fast retransmission of the TCP can be triggered, so that the TCP timeout caused by the order preservation timeout under the dual-link operation is prevented, the transmission rate of the TCP is ensured, and the TCP performance of the multi-link is optimized.
The above-mentioned effect is illustrated by taking the buffer calculation method of the present embodiment as an example applied to a dual link operation (for example, in a dual link download). When the dual uplink works simultaneously, the effect that the network flow at least exceeds the negotiation rate of the main link is expected to be achieved, and the dual link has implementation significance.
For example, FTP (File Transfer Protocol) is used for downloading through a DSL link and an LTE link, where a primary link is the DSL link and a backup link is the LTE link, and when network traffic exceeds a negotiated rate of the DSL link, the network traffic is shunted to the LTE link. The DSL negotiates the rate of 20M, the time delay is 31ms, and the LTE time delay is 60 ms-150 ms.
The method comprises the steps of presetting an FTP server, a client, a DSLAM (Digital Subscriber Line Access Multiplexer), an LTE base station and the like, presetting a route, and ensuring that data can be forwarded in a DSL link and an LTE at the same time.
Downloading through FTP, packet capturing and analyzing TCP messages at an FTP client side, and observing the effect; the packet capture refers to operations such as interception, retransmission, editing, and unloading of data packets transmitted and received by the network, and is used for checking network security or performing data capture.
If no order-preserving method is used, since TCP transmission is a burst flow, the average rate of the time when the upper layer device shunts does not reach DSL negotiation rate 20M, and the FTP client has a large amount of disordered messages, the final rate is stabilized at 5-6M.
By using the order preserving method of the embodiment, the FTP client side packet is not found to have the out-of-order message, and dynamic cache (buffer) changes in real time, and the download rate reaches about 33M.
The method is characterized in that congestion packet loss is simulated on an LTE link through a network loss instrument, a traditional order-preserving method is used, an FTP client side performs packet capture analysis, a small amount of out-of-order messages exist, but the order-preserving process is finished only when the order-preserving Time is up or the order-preserving queue is full due to the fact that the packet loss cannot be predicted by the traditional order-preserving method, even if the small amount of out-of-order triggers rapid TCP retransmission, the RTT (Round-Trip Time) Time of the TCP is prolonged, the transmission rate of the TCP is reduced, and the downloading rate is about 15M.
Simulating congestion packet loss on an LTE link through a network loss instrument, using the dynamic order-preserving method of the embodiment of the scheme, and using an FTP clientAnd (3) the end grabs the packet for analysis, a small amount of out-of-order messages exist, TCP (transmission control protocol) quick retransmission is triggered, and the downloading rate is about 25M. As can be seen from the foregoing, the condition for ending the order preservation in this example, n ≧ DeltaT L -ΔT D ) Δ T, when the buffer in the order-preserving queue exceeds a threshold (Δ T) L -ΔT D ) And when the time is/delta t, the order-preserving process is considered to be finished, so that the buffering time of the TCP in the link is shortened, the RTT is shortened, the rapid retransmission of the TCP is triggered, and the transmission performance of the TCP is improved.
To sum up, in this embodiment, the number of packet buffers is calculated according to the change of the link delay, and the order preserving process is ended when the number of the packets Wen Huancun in the current order preserving queue is greater than the calculated number of the packet buffers, so that whether the order preserving process is ended can be determined according to the calculated threshold and the current number of the buffered packets, and the order preserving process is ended in time when the network packet is lost, thereby preventing TCP from overtime due to order preservation, ensuring the transmission rate of TCP, triggering TCP fast retransmission, and improving the transmission performance of TCP.
Fifth embodiment
Based on the same inventive concept, the embodiment of the present invention further provides a device for cache calculation, and since the principle of the device for solving the problem is similar to the method for cache calculation in fig. 1 to fig. 4 in the embodiment of the present invention, the implementation of the device may refer to the implementation of the method, and the repetition points are not described again.
Referring to fig. 5, a schematic structural diagram of an apparatus for cache calculation is shown, and the apparatus includes an obtaining module 501, a calculating module 502, an updating module 503, and a terminating module 504.
The obtaining module 501 is configured to obtain an average packet time interval of multiple links, a first link delay of an unstable link, and a second link delay of a stable link;
the calculating module 502 is configured to calculate the number of buffered packets according to the first link delay, the second link delay, and the average packet time interval;
the updating module 503 is configured to recalculate the number of buffered packets if the first link delay of the unstable link meets the updating rule;
the termination module 504 is configured to end the order preserving process if the number of the obtained reports Wen Huancun in the current order preserving queue exceeds the calculated number of the message caches.
In this embodiment, the acquisition module 501 includes a counting unit 5011 and a first processing unit 5012.
The counting unit 5011 is configured to obtain the number of messages passing through the multilink within a first preset time by using a sampling method;
the first processing unit 5012 is configured to calculate an average packet time interval according to the first preset time and the number of packets.
In this embodiment, the obtaining module 501 further includes a second processing unit 5013, configured to calculate a second link delay, where the calculation formula is as follows:
ΔT D ＝(ΔT time of return package -ΔT Uplink interleaving -ΔT D downlink interleaving )/2+ΔT D downstream interleaving
Wherein, Δ T D For the second link delay, Δ T Time of return package For the repackaging time, Δ T Uplink interleaving For uplink interleaving time, Δ T D downlink interleaving Is the downlink interleaving time.
In this embodiment, the obtaining module 501 further includes a third processing unit 5014, configured to calculate the first link delay, where the third processing unit 5014 includes:
a first recording subunit: the system comprises a stable link, a sending end and a receiving end, wherein the stable link is used for sending a group of messages to the sending end through the stable link and recording the sending time of the stable link as first time, and the confirmation messages contain expected serial numbers;
a second recording subunit: the second time is recorded after the next group of messages which are sent by the sending end through the unstable link and contain expected serial numbers are received;
a first calculation subunit: and calculating the data flow delay of the unstable link according to the second time, the first time and the uplink delay of the stable link, wherein the calculation formula is as follows:
ΔT L1 ＝T2-T1-ΔT D2
wherein, delta T L1 For data stream delay, T1 is a first time, T2 is a second time, Δ T D2 For stabilizing the linkThe uplink delay of (2);
a second calculation subunit: calculating the average value of the data stream time delay according to the acquired data stream time delay of at least two groups of unstable links, wherein the average value is the first link time delay delta T L 。
In this embodiment, the updating module 503 includes a recalculating unit 5031, configured to recalculate the number of the message caches if, of the obtained multiple first link delays, the latest obtained delay and the last obtained delay meet the updating rule.
In summary, the apparatus of this embodiment obtains the average packet time interval, the link delay, and the link delay difference through the obtaining module, calculates the number of packet caches through the calculating module, and recalculates the number of packet caches by the updating module 503 when the link delay meets the updating rule, so that resources and overhead can be reasonably utilized, and the problem of disorder caused by delays of different links is solved.
It should be appreciated that reference throughout this specification to "one embodiment" or "an embodiment" means that a particular feature, structure or characteristic described in connection with the embodiment is included in at least one embodiment of the present invention. Thus, the appearances of the phrases "in one embodiment" or "in an embodiment" in various places throughout this specification are not necessarily all referring to the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.
In various embodiments of the present invention, it should be understood that the sequence numbers of the above-mentioned processes do not imply an order of execution, and the order of execution of the processes should be determined by their functions and inherent logic, and should not limit the implementation process of the embodiments of the present invention in any way.
In addition, the terms "system" and "network" are often used interchangeably herein.
In the several embodiments provided in the present application, it should be understood that the disclosed method and apparatus may be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the units is only one logical division, and other divisions may be realized in practice, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.
In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be physically included alone, or two or more units may be integrated into one unit. The integrated unit can be realized in a form of hardware, or in a form of hardware plus a software functional unit.
The integrated unit implemented in the form of a software functional unit may be stored in a computer readable storage medium. The software functional unit is stored in a storage medium and includes several instructions to enable a computer device (which may be a personal computer, a server, or a network device) to execute some steps of the transceiving method according to various embodiments of the present invention. And the aforementioned storage medium includes: various media capable of storing program codes, such as a usb disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk.
While the foregoing is directed to the preferred embodiment of the present invention, it will be understood by those skilled in the art that various changes and modifications may be made without departing from the spirit and scope of the invention as defined in the appended claims.
Claims (15)
1. A method of cache calculation for a multilink, said multilink including at least one unstable link and at least one stable link, said method comprising:
acquiring an average packet time interval of the multilink, a first link delay of the unstable link, and a second link delay of the stable link, wherein the acquiring the average packet time interval of the multilink includes: acquiring the number of messages passing through a multilink within a first preset time by a sampling method; calculating an average message time interval according to the first preset time and the number of the messages;
and calculating the number of message caches according to the average message time interval, the first link time delay and the second link time delay.
2. The method of claim 1, wherein the average packet time interval is calculated by the formula:
Δt＝ΔT/m
wherein, Δ T is an average message time interval, Δ T is a first preset time, m is a message number, the first preset time is less than the longest link delay, and the message number is less than the message buffer number.
3. The method of claim 1, wherein the second link delay is calculated by:
ΔT D ＝(ΔT time of return package -ΔT Uplink interleaving -ΔT D downlink interleaving )/2+ΔT D downlink interleaving
Wherein, delta T D For the second link delay, Δ T Time of return package For the time of repacking, Δ T Uplink interleaving For uplink interleaving time, Δ T D downlink interleaving Is the downlink interleaving time.
4. The method of claim 1, wherein the first link delay is calculated by:
after receiving a group of messages sent by a sending end, sending a confirmation message to the sending end by the stable link, and recording the sending time of the stable link as a first time, wherein the confirmation message comprises an expected Sequence Number;
when a next group of messages which are sent by the sending end through an unstable link and contain the expected serial number are received, recording second time;
and calculating the data stream delay of the unstable link according to the second time, the first time and the uplink delay of the stable link, wherein the calculation formula is as follows:
ΔT L1 ＝T2-T1-ΔT D2
wherein, delta T L1 For data stream delay, T1 is a first time, T2 is a second time, Δ T D2 Uplink delay for a stable link;
calculating the average value of the data stream time delay according to the acquired data stream time delay of at least two groups of unstable links, wherein the average value is the first link time delay delta T L 。
5. The method of claim 1, wherein the formula for calculating the number of message buffers is:
n≥(ΔT L -ΔT D )/Δt
wherein n is the number of message caches, and delta T L For the first link delay, Δ T D For the second link delay, Δ t is the average message time interval.
6. The method of claim 1, further comprising:
and if the first link time delay of the unstable link meets the updating rule, recalculating the message cache number.
7. The method of claim 6, wherein the recalculating the number of packet buffers if the first link delay of the unstable link satisfies the update rule comprises:
and if the latest acquired time delay and the last acquired time delay in the acquired time delays of the plurality of first links meet the updating rule, recalculating the number of the message caches.
8. The method of claim 7, wherein the update rule is:
T n ≥T n-1 * Beta or T n ≤T n-1 *β
Wherein, T n For the latest acquired first link delay, T n-1 For the first link delay obtained last time, β is an influence factor, and its value is determined by the first link delay of the unstable link.
9. The method of claim 1, further comprising:
and if the number of the obtained reports Wen Huancun in the current order-preserving queue exceeds the calculated number of the message caches, ending the order-preserving process.
10. An apparatus for cache computing, the apparatus comprising:
an acquisition module: the method comprises the steps of obtaining a multilink average message time interval, a first link time delay of an unstable link and a second link time delay of a stable link; wherein the acquisition module comprises:
a counting unit: the method comprises the steps of obtaining the number of messages passing through a multilink within a first preset time by a sampling method;
a first processing unit: the average message time interval is calculated according to the first preset time and the number of the messages;
a calculation module: and the message cache number is calculated according to the first link time delay, the second link time delay and the average message time interval.
11. The apparatus of claim 10, wherein the obtaining module further comprises a second processing unit configured to calculate a second link delay according to the following formula:
ΔT D ＝(ΔT time of return package -ΔT Uplink interleaving -ΔT D downlink interleaving )/2+ΔT D downlink interleaving
Wherein, delta T D For the second link delay, Δ T Time of return package For the repackaging time, Δ T Uplink interleaving For the uplink interleaving time, Δ T D The downlink interleaving is downlink interleaving time.
12. The apparatus of claim 10, wherein the obtaining module further comprises a third processing unit configured to calculate the first link delay, wherein the third processing unit comprises:
a first recording subunit: the system comprises a stable link, a sending end and a receiving end, wherein the stable link is used for sending a group of messages to the sending end, then the stable link sends a confirmation message to the sending end, and the sending time of the stable link is recorded as first time, wherein the confirmation message contains an expected serial number;
a second recording subunit: the second time is recorded after the next group of messages which are sent by the sending end through the unstable link and contain the expected serial number is received;
a first calculation subunit: and calculating the data stream delay of the unstable link according to the second time, the first time and the uplink delay of the stable link, wherein the calculation formula is as follows:
ΔT L1 ＝T2-T1-ΔT D2
wherein, delta T L1 For data stream delay, T1 is a first time, T2 is a second time, Δ T D2 Uplink delay for a stable link;
a second calculation subunit: calculating the average value of the data stream time delay according to the acquired data stream time delay of at least two groups of unstable links, wherein the average value is the first link time delay delta T L 。
13. The apparatus of claim 10, further comprising:
an update module: and the message cache number is recalculated if the first link delay of the unstable link meets the updating rule.
14. The apparatus of claim 13, wherein the update module comprises:
a recalculation unit: and if the latest acquired time delay and the last acquired time delay in the acquired time delays of the plurality of first links meet the updating rule, recalculating the number of the message caches.
15. The apparatus of claim 10, further comprising:
a termination module: and if the number of the obtained reports Wen Huancun in the current order-preserving queue exceeds the calculated number of the message caches, ending the order-preserving process.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710103163.6A CN108512708B (en) | 2017-02-24 | 2017-02-24 | Cache calculation method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710103163.6A CN108512708B (en) | 2017-02-24 | 2017-02-24 | Cache calculation method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108512708A CN108512708A (en) | 2018-09-07 |
| CN108512708B true CN108512708B (en) | 2023-01-06 |
Family
ID=63372850
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710103163.6A Active CN108512708B (en) | 2017-02-24 | 2017-02-24 | Cache calculation method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108512708B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110944358B (en) | 2018-09-25 | 2021-12-24 | 华为技术有限公司 | Data transmission method and device |
| CN109936514B (en) * | 2019-04-29 | 2022-06-21 | 新华三信息安全技术有限公司 | Message processing method and device |
| CN114024924B (en) * | 2022-01-05 | 2022-04-12 | 北京安博通科技股份有限公司 | TCP stream recombination method and device, electronic equipment and storage medium |
| CN114071840B (en) * | 2022-01-18 | 2022-03-29 | 南京秦之邦科技有限公司 | Remote control system and method for urban lamps |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5430710A (en) * | 1992-03-27 | 1995-07-04 | International Business Machines Corp. | Order preserving transmission of data streams |
| CN1798085A (en) * | 2004-12-28 | 2006-07-05 | 华为技术有限公司 | Method for adjusting size of transmission buffer in control sub layer of wireless link |
| CN1855878A (en) * | 2005-04-28 | 2006-11-01 | 华为技术有限公司 | Voice transmission based on transmission control protocol |
| CN103825689A (en) * | 2014-03-03 | 2014-05-28 | 中国人民解放军理工大学 | Delay determinacy message retransmission method with local cache |
| WO2015196385A1 (en) * | 2014-06-25 | 2015-12-30 | 华为技术有限公司 | Method and device for determining transmission buffer size |
| CN105873096A (en) * | 2016-03-24 | 2016-08-17 | 重庆邮电大学 | Optimization method of efficient throughput capacity of multipath parallel transmission system |
-
2017
- 2017-02-24 CN CN201710103163.6A patent/CN108512708B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5430710A (en) * | 1992-03-27 | 1995-07-04 | International Business Machines Corp. | Order preserving transmission of data streams |
| CN1798085A (en) * | 2004-12-28 | 2006-07-05 | 华为技术有限公司 | Method for adjusting size of transmission buffer in control sub layer of wireless link |
| CN1855878A (en) * | 2005-04-28 | 2006-11-01 | 华为技术有限公司 | Voice transmission based on transmission control protocol |
| CN103825689A (en) * | 2014-03-03 | 2014-05-28 | 中国人民解放军理工大学 | Delay determinacy message retransmission method with local cache |
| WO2015196385A1 (en) * | 2014-06-25 | 2015-12-30 | 华为技术有限公司 | Method and device for determining transmission buffer size |
| CN105873096A (en) * | 2016-03-24 | 2016-08-17 | 重庆邮电大学 | Optimization method of efficient throughput capacity of multipath parallel transmission system |
Non-Patent Citations (2)
| Title |
|---|
| 一种按需分配的多路径传输分组调度算法;曹宇等;《软件学报》;20120715(第07期);全文 * |
| 多通道并行传输中接收缓存阻塞的缓解方法;郑艳伟等;《计算机工程与应用》;20120827(第29期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108512708A (en) | 2018-09-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108512708B (en) | Cache calculation method and device | |
| US10057147B2 (en) | Apparatus and method for controlling data flow in communication system | |
| US11171862B2 (en) | Multi-subflow network transmission method and apparatus | |
| WO2020207479A1 (en) | Method and device for controlling network congestion | |
| JP6274113B2 (en) | Data transmission apparatus, data transmission method, and program thereof | |
| WO2016090978A1 (en) | Traffic switching method and device | |
| CN110445722A (en) | Jamming control method, device, equipment and storage medium | |
| WO2019105317A1 (en) | Method for sending service packet, network device, and system | |
| US8514700B2 (en) | MLPPP occupancy based round robin | |
| US20220103465A1 (en) | Multi-Subflow Network Transmission Method and Apparatus | |
| JP2015076845A (en) | Communication system, control method, and control device | |
| JP4328794B2 (en) | COMMUNICATION SYSTEM, COMMUNICATION DEVICE, AND TRANSMISSION CONTROL METHOD | |
| WO2019244966A1 (en) | Communication device, communication method, and program | |
| JP2017073636A (en) | Relay device and relay method | |
| CA2940077C (en) | Buffer bloat control | |
| Zheng et al. | An effective approach to preventing TCP incast throughput collapse for data center networks | |
| US10855597B2 (en) | Channel coding for real time wireless traffic | |
| EP2002609B1 (en) | Method and termination node for bundling multiple messages into a packet | |
| JPWO2014171543A1 (en) | Data transmission apparatus, data transmission method, and program thereof | |
| CN108200605A (en) | The data transmission method and device of a kind of base station | |
| JP2009105662A (en) | Multi-hop communication system, multi-hop communication method, terminal device and repeating device | |
| Adhari et al. | LEDBAT-MP--On the Application of" Lower-than-Best-Effort" for Concurrent Multipath Transfer | |
| CN108964933B (en) | Method, device and system for transmitting message | |
| JPWO2020012973A1 (en) | Communication controllers, methods, and programs | |
| JP2016201579A (en) | Communication device and transmission band control method |
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 |