CN106375237B - A kind of Hash function Key value screening technique based on 825,990,000 Broadcom of Intel - Google Patents
A kind of Hash function Key value screening technique based on 825,990,000 Broadcom of Intel Download PDFInfo
- Publication number
- CN106375237B CN106375237B CN201610857637.1A CN201610857637A CN106375237B CN 106375237 B CN106375237 B CN 106375237B CN 201610857637 A CN201610857637 A CN 201610857637A CN 106375237 B CN106375237 B CN 106375237B
- Authority
- CN
- China
- Prior art keywords
- secret key
- hash function
- key
- keya
- candidate
- 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
- 238000000034 method Methods 0.000 title claims abstract description 39
- 238000012216 screening Methods 0.000 title claims abstract description 18
- 239000006185 dispersion Substances 0.000 claims abstract description 49
- 238000006073 displacement reaction Methods 0.000 claims abstract description 9
- 230000009467 reduction Effects 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 239000004744 fabric Substances 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 238000001914 filtration Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/12—Avoiding congestion; Recovering from congestion
- H04L47/125—Avoiding congestion; Recovering from congestion by balancing the load, e.g. traffic engineering
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L45/00—Routing or path finding of packets in data switching networks
- H04L45/74—Address processing for routing
- H04L45/745—Address table lookup; Address filtering
- H04L45/7453—Address table lookup; Address filtering using hashing
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Computer And Data Communications (AREA)
Abstract
The invention discloses a kind of data packet shunt method based on 825,990,000 Broadcom of Intel and its Hash function Key value screening technique, which includes: the preparatory arranged distribution code key Secret Key met for obtaining and meeting 40 byte Key values of homologous chummage；Displacement xor operation is carried out to Secret Key by the Hash function built in 825,990,000 Broadcom of Intel, obtains several candidate Secret Key；Obtain 32 valuations to be evaluated of current candidate Secret Key left end after each candidate Secret Key shifts each time；Rear 7 dispersions of 32 valuations to be evaluated of each candidate Secret Key are assessed respectively；Candidate Secret Key corresponding to highest 32 valuations to be evaluated of dispersion is chosen as Hash function Key value.The data packet received from the network port can be made to meet the homologous chummage requirement of upper layer application and make the effect of hardware shunt more balanced, effectively ensure that server overall load is balanced.
Description
Technical field
The present invention relates to server technology fields, more particularly to a kind of data based on 825,990,000 Broadcom of Intel
Packet shunt method and its Hash function Key value screening technique.
Background technique
With the fast development of Internet technology, effect of the network in daily life, study and work is more next
It is heavier.A large amount of, various data traffics are flooded on network link, to ensure data packet be upper layer application used in,
It just needs to handle data packet in bottom by hardware network interface card and corresponding driving, then can just pass to upper layer application.
While the overall performance in order to guarantee server, need network interface card to shunt data packet under normal circumstances, so that connecing from port
The data packet received is distributed evenly over as much as possible in each hardware queue.
Homologous chummage is a demand of the data packet that upper layer application receives network interface card, and basic demand is: identical
Its data packet mutually sent of source and destination passes through network interface card hardware shunt, it should be assigned in the same stream (queue).It is homologous same
Place may insure to apply between data interaction consistency, the upper layer application high for data interaction coherence request has important
Meaning.And be based on the Hash function built in it for 825,990,000 Broadcom hardware shunt of Intel, the Hash function is to receive
Packet five-tuple (protocol type TCP/UDP, source IP address, source port, purpose IP address, destination port, behind 4 can band
Have mask) and 40 byte KEY value as input.Under normal conditions, the KEY value provided according to 82599 handbook of Intel
As one of the input value of Hash function, but the KEY value is not able to satisfy the specific demands such as the homologous chummage of upper layer application,
Shunting effect is also ideal not to the utmost.
Thus, how to obtain more satisfactory KEY value and carry out Hash, the data packet received from the network port can be made full
The homologous chummage requirement of sufficient upper layer application, and the effect of hardware shunt can be made more balanced, effectively ensure that server entirety
Load balancing is a technical problem that technical personnel in the field need to solve at present.
Summary of the invention
The object of the present invention is to provide a kind of data packet shunt method and its Hash based on 825,990,000 Broadcom of Intel
Function Key value screening technique, the homologous chummage that the data packet received from the network port can be made to meet upper layer application are wanted
It asks, and the effect of hardware shunt can be made more balanced, effectively ensure that server overall load is balanced.
In order to solve the above technical problems, the present invention provides the following technical scheme that
A kind of Hash function Key value screening technique based on 825,990,000 Broadcom of Intel, comprising:
The arranged distribution code key Secret Key met for meeting 40 byte Key values of homologous chummage is obtained in advance；
Displacement exclusive or behaviour is carried out to the Secret Key by the Hash function built in 825,990,000 Broadcom of Intel
Make, obtains several candidate Secret Key；
Obtain after each candidate Secret Key shifts each time, 32 of current candidate Secret Key left end to
Assessed value；
Rear 7 dispersions of 32 valuations to be evaluated of each candidate Secret Key are assessed respectively；
Candidate Secret Key corresponding to highest 32 valuations to be evaluated of dispersion is chosen as Hash function Key value.
Preferably, the preparatory acquisition meets the arranged distribution code key of 40 byte Key values of homologous chummage met
Secret Key, are as follows:
The arrangement met point for the 40 byte Key values for meeting homologous chummage being inferred to according to Hash function is obtained in advance
Cloth code key Secret Key:
Secret Key=
KeyA, KeyA, KeyA, KeyB, KeyA,
KeyA, KeyA, KeyB, KeyA, KeyC
,
Wherein, KeyA, KeyB and KeyC are 32 integers of no symbol.
Preferably, the Secret Key is shifted by the Hash function built in 825,990,000 Broadcom of Intel
Xor operation obtains several candidate Secret Key, comprising:
First random number is successively chosen as initial candidate value；
Hash function built in 825,990,000 Broadcom of Intel calculates KeyAA [32], wherein KeyAA [32] is will
(KeyA, KeyA) is moved to left 32 times, moves to left the set of 32 place value of left end every time；
Judge whether calculated current KeyAA [32] is identical with history KeyAA [32]；
If so, giving up current KeyAA [32]；
If it is not, then evaluating the dispersion of rear 7 place value of current KeyAA [32]；
Filter out several KeyA that dispersion is greater than preset threshold；
Choose several KeyB and KeyC for meeting preset condition；
According to described several KeyA, KeyB and the KeyC selected, corresponding candidate Secret Key is obtained.
Preferably, several KeyB and KeyC for meeting preset condition are chosen, comprising:
Second random number is successively chosen as initial candidate value；
Hash function built in 825,990,000 Broadcom of Intel calculates KeyAB [32], wherein KeyAB [32]
To move to left (KeyA, KeyB) 32 times, the set of 32 place value of left end is moved to left every time；
Judge whether calculated current KeyAB [32] is identical with history KeyAB [32]；
If so, giving up current KeyAB [32]；
If it is not, then evaluating the dispersion of rear 7 place value of current KeyAB [32]；
Filter out several KeyB that dispersion is greater than preset threshold.
Preferably, several KeyB and KeyC for meeting preset condition are chosen, further includes:
A third random number is successively chosen as initial candidate value；
Hash function built in 825,990,000 Broadcom of Intel calculates KeyAC [32], wherein KeyAC [32]
To move to left (KeyA, KeyC) 32 times, the set of 32 place value of left end is moved to left every time；
Judge whether calculated current KeyAC [32] is identical with history KeyAC [32]；
If so, giving up current KeyAC [32]；
If it is not, then evaluating the dispersion of rear 7 place value of current KeyAC [32]；
Filter out several KeyC that dispersion is greater than preset threshold.
Preferably, rear 7 dispersions of 32 valuations to be evaluated of each candidate Secret Key are commented respectively
Estimate；Candidate Secret Key corresponding to highest 32 valuations to be evaluated of dispersion is chosen as Hash function Key value, are as follows:
Calculate separately rear 7 variances of 32 valuations to be evaluated of each candidate Secret Key；
The corresponding variance of each candidate Secret Key is compared；
Candidate Secret Key corresponding to maximum variance is chosen as Hash function Key value.
A kind of data packet shunt method based on 825,990,000 Broadcom of Intel, comprising:
Previously according to the Hash function Key value screening as described in any one of the above embodiments based on 825,990,000 Broadcom of Intel
Method filters out Hash function Key value；
Obtain the Hash function Key value and data packet to be shunted；
32 Hash results are calculated by the Hash function of 825,990,000 Broadcom of Intel；
By low 7 values of 32 Hash results, by the allocation of packets to be shunted to 128 reception rings
On；
By preset re-direction table, the data packet to be shunted that 128 receive on ring will be assigned to and be redirected to 16 firmly
In part queue.
Compared with prior art, above-mentioned technical proposal has the advantage that
A kind of Hash function Key value screening technique based on 825,990,000 Broadcom of Intel provided by the present invention, packet
It includes: obtaining the arranged distribution code key Secret Key met for meeting 40 byte Key values of homologous chummage in advance；Pass through
Hash function built in 825,990,000 Broadcom of Intel carries out displacement xor operation to Secret Key, obtains several candidates
Secret Key；Obtain after each candidate Secret Key shifts each time, 32 of current candidate Secret Key left end to
Assessed value；Rear 7 dispersions of 32 valuations to be evaluated of each candidate Secret Key are assessed respectively；Choose dispersion
Candidate Secret Key corresponding to highest 32 valuations to be evaluated is as Hash function Key value.In technical solution of the present invention
In, the arranged distribution code key Secret Key met for meeting 40 byte Key values of homologous chummage is obtained first, is then passed through
Hash function calculates the Secret Key, and obtains several candidate Secret Key, chooses dispersion highest 32
Candidate Secret Key corresponding to valuation to be evaluated is as Hash function Key value, since 825,990,000 Broadcom of Intel is root
Hash result is calculated according to Hash function Key value to be shunted to treat streamed data packet, and in order to guarantee homologous chummage and hard
Part shunting effect is more balanced, then is necessarily required to shunt band streamed data packet with higher dispersion, candidate Secret
Key dispersion is bigger, then the Hash result being calculated is more balanced to the shunting effect with streamed data packet, it is thus possible to most
The equilibrium of the shunting of streamed data packet is treated in guarantee in big degree.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is the present invention
Some embodiments for those of ordinary skill in the art without creative efforts, can also basis
These attached drawings obtain other attached drawings.
Fig. 1 is the Hash function based on 825,990,000 Broadcom of Intel provided by a kind of specific embodiment of the present invention
Key value screening technique flow chart；
Fig. 2 is the data packet shunting side based on 825,990,000 Broadcom of Intel provided by one embodiment of the present invention
Method flow chart.
Specific embodiment
Core of the invention is to provide a kind of data packet shunt method and its Hash based on 825,990,000 Broadcom of Intel
Function Key value screening technique, the homologous chummage that the data packet received from the network port can be made to meet upper layer application are wanted
It asks, and the effect of hardware shunt can be made more balanced, effectively ensure that server overall load is balanced.
It is understandable in order to enable the above objects, features and advantages of the present invention to become apparent, with reference to the accompanying drawing to this hair
Bright specific embodiment is described in detail.
Detail is elaborated in the following description to fully understand the present invention.But the present invention can with it is a variety of not
Other way described herein is same as to implement, those skilled in the art can do class without violating the connotation of the present invention
Like popularization.Therefore the present invention is not limited by following public specific embodiment.
Referring to FIG. 1, Fig. 1 is provided by a kind of specific embodiment of the present invention based on ten thousand Broadcom of Intel82599
Hash function Key value screening technique flow chart.
A kind of specific embodiment of the invention provides a kind of Hash function based on 825,990,000 Broadcom of Intel
Key value screening technique, comprising:
S11: the arranged distribution code key Secret met for meeting 40 byte Key values of homologous chummage is obtained in advance
Key。
In one embodiment of the invention, being met for the 40 byte Key values for meeting homologous chummage is obtained in advance
Arranged distribution code key Secret Key, are as follows: the 40 byte Key for meeting homologous chummage being inferred to according to Hash function are obtained in advance
The arranged distribution code key Secret Key of value met.
Inventor requires the study found that meet homologous chummage of the IPV4 and IPV6 message about source IP and destination IP
Message Hash result between source IP and destination IP is consistent, i.e., after source IP and purpose IP address exchange, the Hash of the two
As a result it is consistent.It is inferred to meet 40 bytes of homologous chummage according to hash algorithm built-in in ten thousand Broadcom of Intel82599
The arranged distribution that Key value meets is as follows:
Secret Key=
KeyA, KeyA, KeyA, KeyB, KeyA,
KeyA, KeyA, KeyB, KeyA, KeyC
,
Wherein, KeyA, KeyB and KeyC are 32 integers of no symbol.
Since a byte is 8, the Key value of 40 bytes totally 320, and KeyA, KeyB and KeyC are no symbol
Number 32 integers, therefore, Secret Key includes 10 above-mentioned 32 integers.
S12: displacement exclusive or behaviour is carried out to Secret Key by the Hash function built in 825,990,000 Broadcom of Intel
Make, obtains several candidate Secret Key；
S13: obtaining after each candidate Secret Key shifts each time, 32 of current candidate Secret Key left end to
Assessed value.Shifting function successively is carried out to each candidate Secret Key, and it is each to obtain current candidate Secret Key
After secondary displacement, 32 valuations to be evaluated of left end, until having obtained after each candidate Secret Key is shifted each time itself
32 valuations to be evaluated of left end.
S14: rear 7 dispersions of 32 valuations to be evaluated of each candidate Secret Key are assessed respectively.
S15: candidate Secret Key corresponding to highest 32 valuations to be evaluated of dispersion is chosen as Hash function Key
Value.
In the present embodiment, the arranged distribution met for obtaining the 40 byte Key values for meeting homologous chummage first is secret
Then key Secret Key calculates the Secret Key by Hash function, several candidate Secret Key is obtained,
And choose candidate Secret Key corresponding to highest 32 valuations to be evaluated of dispersion as Hash function Key value, due to
Ten thousand Broadcom of Intel82599 is to calculate Hash result according to Hash function Key value to be shunted to treat streamed data packet,
And in order to guarantee homologous chummage and hardware shunt effect is more balanced, then it is necessarily required to higher dispersion to band streamed data
Packet is shunted, and candidate Secret Key dispersion is bigger, then the Hash result being calculated is to the shunting with streamed data packet
Effect is more balanced, it is thus possible to guarantee to treat the equilibrium of the shunting of streamed data packet to the full extent.
In one embodiment of the invention, by 825,990,000 Broadcom of Intel built in Hash function pair
Secret Key carries out displacement xor operation, obtains several candidate Secret Key, comprising:
First random number is successively chosen as initial candidate value；According to built in 825,990,000 Broadcom of Intel
Hash function calculates KeyAA [32], wherein KeyAA [32] is to move to left (KeyA, KeyA) 32 times, is moved to left every time most left
Hold the set of 32 place values；Judge whether calculated current KeyAA [32] is identical with history KeyAA [32]；Work as if so, giving up
Preceding KeyAA [32]；If it is not, then evaluating the dispersion of rear 7 place value of current KeyAA [32]；Dispersion is filtered out to be greater than
Several KeyA of preset threshold；Choose several KeyB and KeyC for meeting preset condition；According to several KeyA, the KeyB selected and
KeyC obtains corresponding candidate Secret Key.
Since the IP of Hash input is arbitrariness, in order to guarantee the discreteness of final Hash result, first is random
Number is randomly generated, using it as the candidate of KeyA, and by above-mentioned judgment method whether to determine first random number
Retain.Wherein, the first random number refers to the candidate random number as KeyA, can choose according to actual needs it is multiple not
Same random number is successively used as the first random number, to obtain several KeyA.
Wherein, several KeyB and KeyC for meeting preset condition are chosen, comprising: successively choose a second random number conduct
Initial candidate value；Hash function built in 825,990,000 Broadcom of Intel calculates KeyAB [32], wherein KeyAB
[32] to move to left (KeyA, KeyB) 32 times, the set of 32 place value of left end is moved to left every time；Judge calculated current KeyAB
[32] whether identical with history KeyAB [32]；If so, giving up current KeyAB [32]；If it is not, then to current KeyAB's [32]
The dispersion of 7 place values is evaluated afterwards；Filter out several KeyB that dispersion is greater than preset threshold.
It should be noted that be at this time also to randomly select the second random number as initial candidate KeyB, but initial candidate
KeyB is cannot be identical with initial candidate KeyA, that is to say, that arbitrary second random number cannot be with any first random number phase
Together, while in order to ensure final Hash as a result, the number of reduction iteration, can examine or check above-mentioned KeyAA [32] 7 afterwards, then
One and 7 initial candidate KeyB being different from after 32 KeyAA is randomly generated.
A third random number is successively chosen as initial candidate value；According to built in 825,990,000 Broadcom of Intel
Hash function calculates KeyAC [32], wherein KeyAC [32] is to move to left (KeyA, KeyC) 32 times, is moved to left every time most left
Hold the set of 32 place values；Judge whether calculated current KeyAC [32] is identical with history KeyAC [32]；Work as if so, giving up
Preceding KeyAC [32]；If it is not, then evaluating the dispersion of rear 7 place value of current KeyAC [32]；Dispersion is filtered out to be greater than
Several KeyC of preset threshold.
It is similar with KeyB is chosen, at this time and third random number is randomly selected as initial candidate KeyC, but initial candidate
KeyC is cannot be identical with initial candidate KeyA, that is to say, that arbitrary third random number cannot be with any first random number phase
Together, while in order to ensure final Hash as a result, the number of reduction iteration, can examine or check above-mentioned KeyAA [32] 7 afterwards, then
One and 7 initial candidate KeyC being different from after 32 KeyAA is randomly generated.
For inventor the study found that due to Hash the result is that one without 32 integers of symbol, hash algorithm is by right
Secret Key carries out what displacement xor operation was realized, and after Secret Key is moved to left each time, 32 place values of left end be will affect
Hash is as a result, dispersion degree will determine the dispersion degree of Hash result entirety.In the present embodiment, it is indicated with KeyAA [32]
(KeyA, KeyA) is moved to left 32 times, moves to left the set of rear 32 place value of Secret Key left end each time；KeyAB [32] is indicated
(KeyA, KeyB) is moved to left 32 times, moves to left the set of rear 32 place value of Secret Key left end each time；KeyBA [32] is indicated
(KeyB, KeyA) moves to left 32 times, each time the set of 32 place value of left end；KeyAC [32] indicates that (KeyA, KeyC) moves to left 32
It is secondary, the set of rear 32 place value of Secret Key left end is moved to left each time.KeyAA[32],KeyAB[32],KeyBA[32],
The dispersion degree of KeyAC [32] determines the dispersion degree of entire Hash result.
When calculating Secret Key, it can be calculated according to the sequence of KeyA, KeyB, KeyC.For calculating KeyA,
A random number is chosen first as initial candidate value, is calculated KeyAA [32], since KeyAA [32] indicate that (KeyA, KeyA) is left
It moves 32 times, moves to left the set of rear 32 place value of Secret Key left end each time, therefore, KeyAA [32] has 32 values, if 32
A KeyAA [32] has identical, then gives up the identical value, i.e., as described above " judges that calculated current KeyAA [32] is
No and history KeyAA [32] are identical；If so, giving up current KeyAA [32] ", this is because if 32 KeyAA [32] have phase
Together, then it represents that reduction dispersion is had, this is because Hash result will in circulation each time in entire Hash calculating process
Decide whether to carry out XOR operation with KeyAA, KeyAB, KeyBA, KeyAC, ideally join each time according to the position of input
With KeyAA, KeyAB, KeyBA of XOR operation, KeyAC should not be identical, hash result step-by-step finally obtained in this way point
Cloth can be random, and discretization degree reaches maximum.For example, if 32 KeyAA have it is identical, identical KeyAA 7 thereafter
And it is identical, according to the characteristic of XOR operation, two identical KeyAA exclusive or results are 0, and any number and 0 exclusive or value are constant,
The two identical KeyAA itself hash is then equivalent to balance out.Otherwise, it is chosen by evaluating the dispersion of 7 place values thereafter
Several candidate KeyA are selected, i.e., according to above-mentioned algorithm, screen the set of one group of mutually different 32 KeyAA [32], and
The set for filtering out mutually different 32 KeyAA [32] of multiple groups in the same manner, can thus filter out several candidates
KeyA。
It is similar with KeyA is calculated, if by the above method of present embodiment filter out respectively it is several to candidate KeyB and
The dry KeyC to candidate.Then, using Secret Key composed by each candidate KeyA, candidate KeyB and candidate KeyC as candidate
Secret Key。
On the basis of any of the above-described embodiment, in one embodiment of the present invention, respectively to each candidate Secret
Rear 7 dispersions of 32 valuations to be evaluated of Key are assessed；It chooses corresponding to highest 32 valuations to be evaluated of dispersion
Candidate Secret Key as Hash function Key value, are as follows: after 32 valuations to be evaluated for calculating separately each candidate Secret Key
7 variances；The corresponding variance of each candidate Secret Key is compared；Choose candidate Secret corresponding to maximum variance
Key is as Hash function Key value.
In the present embodiment, the appraisal procedure of the dispersion of each candidate Secret Key is illustrated.Each candidate
Rear 7 dispersions of 32 place values of left end decide the discrete journey of Hash result entirety after Secret Key is shifted each time
Degree.In the present embodiment, rear 7 variances of 32 valuations to be evaluated of each candidate Secret Key are calculated separately, variance is got over
It is big then indicate that dispersion is higher, thus select it is final required for candidate Secret Key as Hash function Key value.
Referring to FIG. 2, Fig. 2 is the number based on 825,990,000 Broadcom of Intel provided by one embodiment of the present invention
According to packet shunt method flow chart.
Correspondingly, one embodiment of the present invention additionally provides a kind of data packet based on 825,990,000 Broadcom of Intel
Shunt method, comprising:
S21: previously according to the Hash letter based on 825,990,000 Broadcom of Intel as provided by above-mentioned any embodiment
Number Key value screening technique, filters out Hash function Key value.
S22: Hash function Key value and data packet to be shunted are obtained.
S23: 32 Hash results are calculated by the Hash function of 825,990,000 Broadcom of Intel.
S24:, will be in allocation of packets to be shunted to 128 reception rings by low 7 values of 32 Hash results.
S25: by preset re-direction table, the data packet to be shunted that will be assigned on 128 reception rings is redirected to 16
In a hardware queue.
In the present embodiment, it first passes through and is as above stated provided by any embodiment based on Intel 825,990,000 million in advance
The Hash function Key value screening technique of network interface card, filters out Hash function Key value.Can make the IPV4 received from network interface card and
IPV6 message meets based on the requirement of the homologous chummage of source IP and destination IP, while making the effect of hardware shunt more balanced, from
And can effectively ensure that server overall load is balanced, hardware foundation is established for the performance of upper layer application function.
In conclusion a kind of data packet based on 825,990,000 Broadcom of Intel point provided by embodiment of the present invention
Stream method and its Hash function Key value screening technique can be such that IPV4 the and IPV6 message received from network interface card meets based on source
IP and the requirement of the homologous chummage of destination IP, while making the effect of hardware shunt more balanced, so as to effectively ensure to service
Device overall load is balanced, establishes hardware foundation for the performance of upper layer application function.
Above to a kind of data packet shunt method and its Hash based on 825,990,000 Broadcom of Intel provided by the present invention
Function Key value screening technique is described in detail.Specific case used herein is to the principle of the present invention and embodiment
It is expounded, the above embodiments are only used to help understand the present invention and its core concept.It should be pointed out that for this
For the those of ordinary skill of technical field, without departing from the principle of the present invention, the present invention can also be carried out several
Improvement and modification, these improvements and modifications also fall within the scope of protection of the claims of the present invention.
Claims (7)
1. a kind of Hash function Key value screening technique based on 825,990,000 Broadcom of Intel characterized by comprising
The arranged distribution code key Secret Key being inferred to according to hash algorithm is obtained in advance；Wherein, the arranged distribution code key
Secret Key is the 40 byte Key values for meeting homologous chummage；
Displacement xor operation is carried out to the Secret Key by the Hash function built in 825,990,000 Broadcom of Intel, is obtained
Take several candidate Secret Key；
It obtains after each candidate Secret Key shifts each time, 32 of current candidate Secret Key left end are to be assessed
Value；
Rear 7 dispersions of 32 valuations to be evaluated of each candidate Secret Key are assessed respectively；
Candidate Secret Key corresponding to highest 32 valuations to be evaluated of dispersion is chosen as Hash function Key value.
2. the method according to claim 1, wherein the arrangement that the preparatory acquisition is inferred to according to hash algorithm
It is distributed code key Secret Key, are as follows:
The arranged distribution met for obtaining the 40 byte Key values for meeting homologous chummage being inferred to according to Hash function in advance is secret
Key Secret Key:
Secret Key=
KeyA, KeyA, KeyA, KeyB, KeyA,
KeyA, KeyA, KeyB, KeyA, KeyC
,
Wherein, KeyA, KeyB and KeyC are 32 integers of no symbol.
3. according to the method described in claim 2, it is characterized in that, passing through the Hash letter built in 825,990,000 Broadcom of Intel
It is several that displacement xor operation is carried out to the Secret Key, obtain several candidate Secret Key, comprising:
First random number is successively chosen as initial candidate value；
Hash function built in 825,990,000 Broadcom of Intel calculates KeyAA [32], wherein KeyAA [32] is will
(KeyA, KeyA) is moved to left 32 times, moves to left the set of 32 place value of left end every time；
Judge whether calculated current KeyAA [32] is identical with history KeyAA [32]；
If so, giving up current KeyAA [32]；
If it is not, then evaluating the dispersion of rear 7 place value of current KeyAA [32]；
Filter out several KeyA that dispersion is greater than preset threshold；
Choose several KeyB and KeyC for meeting preset condition；
According to described several KeyA, KeyB and the KeyC selected, corresponding candidate Secret Key is obtained.
4. according to the method described in claim 3, it is characterized in that, selection meets several KeyB and KeyC of preset condition, packet
It includes:
Second random number is successively chosen as initial candidate value；
Hash function built in 825,990,000 Broadcom of Intel calculates KeyAB [32], wherein KeyAB [32] is will
(KeyA, KeyB) is moved to left 32 times, moves to left the set of 32 place value of left end every time；
Judge whether calculated current KeyAB [32] is identical with history KeyAB [32]；
If so, giving up current KeyAB [32]；
If it is not, then evaluating the dispersion of rear 7 place value of current KeyAB [32]；
Filter out several KeyB that dispersion is greater than preset threshold.
5. according to the method described in claim 4, it is characterized in that, choosing several KeyB and KeyC for meeting preset condition, also
Include:
A third random number is successively chosen as initial candidate value；
Hash function built in 825,990,000 Broadcom of Intel calculates KeyAC [32], wherein KeyAC [32] is will
(KeyA, KeyC) is moved to left 32 times, moves to left the set of 32 place value of left end every time；
Judge whether calculated current KeyAC [32] is identical with history KeyAC [32]；
If so, giving up current KeyAC [32]；
If it is not, then evaluating the dispersion of rear 7 place value of current KeyAC [32]；
Filter out several KeyC that dispersion is greater than preset threshold.
6. method according to any one of claims 1 to 5, which is characterized in that respectively to each candidate Secret Key
Rear 7 dispersions of 32 valuations to be evaluated assessed；Choose time corresponding to highest 32 valuations to be evaluated of dispersion
Select Secret Key as Hash function Key value, are as follows:
Calculate separately rear 7 variances of 32 valuations to be evaluated of each candidate Secret Key；
The corresponding variance of each candidate Secret Key is compared；
Candidate Secret Key corresponding to maximum variance is chosen as Hash function Key value.
7. a kind of data packet shunt method based on 825,990,000 Broadcom of Intel characterized by comprising
Previously according to the Hash function Key value such as claimed in any one of claims 1 to 6 based on 825,990,000 Broadcom of Intel
Screening technique filters out Hash function Key value；
Obtain the Hash function Key value and data packet to be shunted；
32 Hash results are calculated by the Hash function of 825,990,000 Broadcom of Intel；
It, will be in the allocation of packets to be shunted to 128 reception rings by low 7 values of 32 Hash results；
By preset re-direction table, the data packet to be shunted that will be assigned on 128 reception rings is redirected to 16 hardware teams
In column.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610857637.1A CN106375237B (en) | 2016-09-28 | 2016-09-28 | A kind of Hash function Key value screening technique based on 825,990,000 Broadcom of Intel |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610857637.1A CN106375237B (en) | 2016-09-28 | 2016-09-28 | A kind of Hash function Key value screening technique based on 825,990,000 Broadcom of Intel |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106375237A CN106375237A (en) | 2017-02-01 |
| CN106375237B true CN106375237B (en) | 2019-11-01 |
Family
ID=57898518
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610857637.1A Active CN106375237B (en) | 2016-09-28 | 2016-09-28 | A kind of Hash function Key value screening technique based on 825,990,000 Broadcom of Intel |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106375237B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107046503B (en) * | 2017-04-24 | 2020-08-04 | 新华三技术有限公司 | Message transmission method, system and device |
| CN112260960A (en) * | 2020-10-15 | 2021-01-22 | 东莞飞思凌通信技术有限公司 | Cross-device multi-link global homology and co-homing implementation method |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102497323A (en) * | 2011-12-19 | 2012-06-13 | 曙光信息产业(北京)有限公司 | System and method used for realizing homologous and homoclinic message division based on intel network interface card |
| CN102523164A (en) * | 2011-12-19 | 2012-06-27 | 曙光信息产业(北京)有限公司 | System and method for implementing complex homologous and homoclinic flow division in network card |
| CN104468412A (en) * | 2014-12-04 | 2015-03-25 | 东软集团股份有限公司 | RSS-based network session data packet distribution method and system |
| CN105871725A (en) * | 2015-01-23 | 2016-08-17 | 阿里巴巴集团控股有限公司 | Message shunt method and device |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7765405B2 (en) * | 2005-02-25 | 2010-07-27 | Microsoft Corporation | Receive side scaling with cryptographically secure hashing |
-
2016
- 2016-09-28 CN CN201610857637.1A patent/CN106375237B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102497323A (en) * | 2011-12-19 | 2012-06-13 | 曙光信息产业(北京)有限公司 | System and method used for realizing homologous and homoclinic message division based on intel network interface card |
| CN102523164A (en) * | 2011-12-19 | 2012-06-27 | 曙光信息产业(北京)有限公司 | System and method for implementing complex homologous and homoclinic flow division in network card |
| CN104468412A (en) * | 2014-12-04 | 2015-03-25 | 东软集团股份有限公司 | RSS-based network session data packet distribution method and system |
| CN105871725A (en) * | 2015-01-23 | 2016-08-17 | 阿里巴巴集团控股有限公司 | Message shunt method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106375237A (en) | 2017-02-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107438994B (en) | Method, apparatus, and computer storage medium for server load balancing | |
| US10298502B2 (en) | Method, device, and system for performing balance adjustment on egress traffic of SDN based idc network | |
| CN106664320B (en) | Mechanism to support a traffic chain graph in a communication network | |
| US9237091B2 (en) | System and method of load balancing for ethernet link aggregation | |
| Basescu et al. | SIBRA: Scalable internet bandwidth reservation architecture | |
| CN108880999B (en) | Method, equipment and system for service routing | |
| CN104145451B (en) | Message processing method, deep-packet detection request network element and deep packet inspection device | |
| JP6348983B2 (en) | Layer 2 path tracking through context coding in software defined networking | |
| CN101127649B (en) | A method and system for preventing from network attacks | |
| CN104811326A (en) | Service chain management method, service chain management system, and devices | |
| WO2016142774A1 (en) | Method and apparatus for load balancing in network switches | |
| US10243857B1 (en) | Method and apparatus for multipath group updates | |
| CN107113235B (en) | Data routing method and device | |
| TW201215063A (en) | Automated traffic engineering for multi-protocol label switching (MPLS) with link utilization as feedback into the tie-breaking mechanism | |
| CN106534394B (en) | Apparatus, system, and method for managing ports | |
| CN101997826A (en) | Routing methods of control net element, forwarding net element and internet protocol network | |
| CN107800668B (en) | Distributed denial of service attack defense method, device and system | |
| WO2010132061A1 (en) | A method and apparatus for policy enforcement using a tag | |
| CN104486236B (en) | The method and routing device of load balancing | |
| CN106375237B (en) | A kind of Hash function Key value screening technique based on 825,990,000 Broadcom of Intel | |
| CN108809847A (en) | Realize the method, apparatus and network system of load balancing | |
| JP2007288711A (en) | Gateway apparatus, setting controller, and load distribution method and program for gateway apparatus | |
| Kim et al. | Towards peer-to-peer traffic analysis using flows | |
| US11750518B2 (en) | Elastic modification of application instances in a network visibility infrastructure | |
| Detal et al. | Revisiting flow-based load balancing: Stateless path selection in data center networks |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | 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 |