WO2013010874A1

WO2013010874A1 - Method and apparatus for synchronizing processes which perform multipoint measurements in a data transmission network

Info

Publication number: WO2013010874A1
Application number: PCT/EP2012/063572
Authority: WO
Inventors: Tanja Zseby; Michael Kleis
Original assignee: Fraunhofer-Gesellschaft zur Förderung der angewandten Forschung e.V.
Priority date: 2011-07-15
Filing date: 2012-07-11
Publication date: 2013-01-24
Also published as: DE102011079279A1; EP2567515A1

Abstract

The invention relates to a method for synchronizing multipoint measurements in a data transmission network (100) having observation apparatuses (11, 12, 13) which perform measurements, wherein the observation apparatuses (11, 12, 13) have associated pulse-coupled oscillator apparatuses (21, 22, 23) which each have a charging function v(t) which increases between v(t=0)=0 and v (t=t_Max) = v_Max, and each of the pulse-coupled oscillator apparatuses (21, 22, 23), upon reaching the threshold value vMax, automatically sends a signal (30) for other pulse-coupled oscillator apparatuses (21, 22, 23) and is then set to v=0, wherein the time interval 0 to t_Maxdetermines the frequencies of the oscillator apparatuses (21, 22, 23), the pulse-coupled oscillator apparatuses (21, 22, 23) which receive the signal (30) involve the charging function v(t) being raised by the absolute value s, with the result that all the oscillator apparatuses (21, 22, 23) which are coupled to one another by signals (30), and hence also the coupled observation apparatuses (11, 12, 13), synchronously perform measurements on the data packets in the data transmission network on the basis of the self-synchronizing properties of the pulse-coupled oscillator apparatuses (21, 22, 23), with a synchronized selection probability p_s of the observation apparatuses (11, 12, 13) being formed for the data traffic from the functional relationship between the charging function v(t) and the selection probability such that the selection probability p_s negatively correlates to the frequency of the oscillator apparatus (21, 22, 23). The invention also relates to an apparatus.

Description

Method and device for synchronizing

Processes that multipoint measurements in one

Perform data transmission network.

The invention relates to a method for synchronizing multipoint measurements in a data transmission network and to a device for synchronizing

Multipoint measurements in a data transmission network.

In the measurement of transmission properties in the

Transmission of data packets (e.g., network delays, packet loss, paths that take packets in the network) in a communications network will be different

Method used. In the case of methods in which the existing data traffic is used for the measurement, the resources required for the measurement are determined according to the currently occurring data traffic

Data transmission network. The data traffic is usually subject to large dynamic fluctuations, so that the necessary resources in the data transmission network

vary. In order to control resource consumption (e.g., not to exceed a set threshold), resource consumption can be reduced through the use of

Sampling to be adapted to the existing resources during the measurement.

A special case are multipoint measurements, i. the measurements are made with at least two

Observation devices. Such measurements are useful to e.g. determine the path of a data packet or the quality of the data transmission in the data transmission network. In such multipoint measurements, the multiplicity of

Observation devices are coordinated to

ensure that the same data package of all

Observation devices is detected.

One possible solution is to use a packet selection method that makes the selection based on a hash value over the packet content. This is on all at

Observation point packets applied a hash function and thus calculated a hash value for each packet. All packets whose hash value falls within a predefined range (hash selection area) are selected, the other packets are discarded. This corresponds to the function of a filter.

Such a method (hash-based packet selection) is u. a. described in RFC 5475. The selection probability can be changed by changing the selection range (e.g., more packages are usually selected for a larger selection range). However, the number of selected packages also depends on the data traffic and can therefore not be defined exactly.

However, to capture the same data packets at all points, this requires a synchronized parameterization of the

Observation devices ahead.

It is the object of the present invention to provide efficient methods and apparatuses with which a synchronization of processes for multipoint measurements in data transmission networks is possible.

The object is achieved by a method having the features of claim 1.

For this purpose, a synchronization of the processes

pulse coupled oscillator devices proposed.

Specifically, the synchronization of the

Selection probabilities (i.e., probability with which a packet is selected for measurement) for the

Sampling processes at the different observation points considered. The selection probability is called

Input used for one or more sampling process (s) at the observation point.

Thereby take at least two observation devices

Measurements in the data transmission network. Result of the measurements can e.g. the path taken by the packet (determination of the sequence of measurement points on which

the same package was observed) or the term that the Package on the way has learned (determination, ie measurement of the difference of arrival times at the observation points).

To control the sampling processes (especially the

Auswahlwahrscheichichkeiten) are the at least two

Observation devices associated with at least two pulse-coupled oscillator devices.

It should be ensured that to the

Observation points in each case the same packages through the

Filters are selected.

The at least two pulse-coupled oscillator devices each have a charging function v (t) which increases from v (t = 0) = 0 to v (t = t _max ) = v _Max . For example, t _Max and v _Max may differ at the different observation points and may change dynamically (eg, depending on the current resource consumption at the observation point).

Each of the pulse coupled oscillator devices automatically sends a signal to other pulse coupled oscillator devices upon reaching the threshold v _Max . After sending the signal, the charging function is set to v = 0 and a new cycle begins. The time interval 0 to t _Max determines the frequencies of the oscillator devices. The faster the reset to v = 0, the higher the frequency of the oscillator devices.

At the signal receiving pulse coupled

Oscillator devices, the charge function v (t) is raised by the amount ε, i. the charging function is the

respective threshold v _max somewhat closer.

Since each oscillator periodically after reaching V _Max

Signals are sent, the coupled

Oscillator devices due to

self-synchronizing properties of pulse-coupled oscillator devices excited to a synchronous oscillation.

Approach of the invention is, using this effect, the parameters for the measurement (eg Selection probability) at the different observation points.

In addition to the described self-synchronizing

Characteristics of pulse-coupled oscillators, can

advantageously be used the fact that with a suitable choice of the oscillator parameters can be achieved that the oscillator with the smallest threshold V _max (corresponds to the highest fire speed) dominates all other oscillators. That is, its firing frequency is transmitted to all oscillator devices.

The procedure is set up so that a

synchronized selection probability of

Observation devices for the traffic from a functional relationship of the charging function v (t) and the selection probability is formed so that the

Selection probability negative with the frequency of

Oscillator device (i.e., when the

As the oscillator frequency increases, the

Selection probability smaller). When the maximum value v _{Max of} the charging function is reduced, the frequency of the oscillator device increases.

In an advantageous embodiment, an association between the loading function and the selection probability is chosen such that the functional relationship between the loading function v (t) and the selection probability is formed such that the new threshold 't ax correlates positively with the new selection probability (ie If the value for the new threshold v'Max is reduced, the new selection probability for the

Reduced data packets in the data transmission network). Thus, the probability of selection can be reduced efficiently if the data volume in the data transmission network becomes too high.

In a particularly advantageous embodiment, the charging function v (t) is strictly monotonically increasing and concave. For a simple calculation of dependencies between the new threshold v 'ax and the new one

Selection probability p ' _s , it is advantageous if in one embodiment, the charging function v (t) of the

Oscillator devices is at least partially linearized. It is particularly advantageous if the relationship between the new threshold v 'ax and the new

Selection probability p ' _s through

a is given (for a> 0).

In an advantageous embodiment of the method, the change of the parameters causes a better utilization of the available resources. This can e.g. the reduction of the selection probability to be an overload of the

Computing capacity or an increase in the

Selection probability to use available computing capacity. In an advantageous embodiment, the at least one observation device reacts adaptively to a change in the data transmission in the

Data transmission network.

In a particularly advantageous embodiment, the pulse-coupled oscillator devices of

Observation devices are essentially the same

Dynamics, i. the charging functions are identical (or essentially identical) and are in the

Data transmission network coupled together, in particular, they exchange signals with each other.

In an advantageous embodiment, the

Synchronization of the selection process of the packages under

Use a hash function. The selection of packets is effected by means of a hash selection area which has the function of a filter. The task is also performed by a device with the

Characteristics of claim 10 solved.

The device has at least two

Observation devices for performing measurements on data packets at intervals, wherein the at least two observation devices are associated with at least two pulse-coupled oscillator devices.

The at least two pulse-coupled oscillator devices each have a charge function v (t) which increases from v (t = 0) = 0 to v (t = t _max ) = v _Max , each of the pulse-coupled oscillator devices automatically reaching the threshold value v _Max a signal for other pulse-coupled

Oscillator devices and then set to v = 0, wherein the time interval 0 to t _{Max determines} the frequencies of the oscillator devices.

At the signal receiving pulse coupled

Oscillator devices, the charging function v (t) is raised by the amount ε.

Such a device shows a self-synchronizing behavior.

In this case, a synchronized selection probability of the traffic observation devices is formed from the functional relationship between the loading function v (t) and the selection probability such that the

Measuring frequency negative with the frequency of

Oscillator device correlated. The device thus has a certain, preset link between the oscillator device frequency and the measurement frequency.

In an advantageous embodiment of the device, the functional relationship between the charging function v (t) is formed so that the new threshold 't ax correlates positively with the new measurement frequency.

In a further advantageous embodiment, the charging function v (t) is strictly monotonically increasing and concave. Also embodiments are advantageous in which the

Charging function v (t) of the oscillator devices is at least partially linearized to a relationship between a new threshold v ' _Max and a new

Selection probability p ' _s to find. A particularly simple relationship arises when the relationship between the new threshold v ' _Max and the new one

Selection probability p ' _s through

a

(for a> 0) is given.

Advantageously, in embodiments all

pulse coupled oscillator devices of

Observation devices are essentially the same

Dynamics and are in the data transmission network

coupled together so that they exchange signals in particular.

The invention will be described below with reference to

Embodiments described. It shows

Fig. 1 is a schematic representation of a

Data transmission network with observation devices for data transmission;

Fig. 2A is an illustration of a charging function for a

Embodiment of a pulse-coupled oscillator device;

Fig. 2B is an illustration of the charging function with an increase in the state variable upon receipt of a signal in one embodiment;

3 is a representation of the functional relationship for determining a new threshold value in a

embodiment;

Fig. 4 is a schematic representation of a

Data transmission network with an embodiment. Data transmission networks 100 are found in all

Areas of the art, such as in telecommunications or networking of computers. The Internet is an example of a communications network 100.

For complex networks, measurements are made simultaneously with multiple observers 11, 12, 13 to track the path of data packets 30 in the network. The synchronization of the observation devices 11, 12, 13 is an object of the present invention.

FIG. 1 schematically shows a data transmission network 100 with three in order to illustrate the basic terms

Observation devices 11, 12, 13 shown. The

Observation devices 11, 12, 13 can be computers or microchips that determine the data volume

Detect points in the data transmission network 100, in particular the data volume per time measure.

If the data volume increases in one area of the data transmission network 100, then it makes sense to use the

To lower the selection probability p _s with which the

Observation devices 11, 12, 13 the data flow

to capture. Since the observation devices 11, 12, 13 can be spatially widely distributed, it is advantageous if the change of the selection probability p _s takes place adaptively and spreads efficiently in the data transmission network 100.

The following describes methods and devices with which an efficient synchronization of the

Observation devices 11, 12, 13 is possible. To

Complete the synchronization of all

Observation devices 11, 12, 13 in

Data processing network 100 the same (or at least substantially the same) selection probability p _s on.

For this purpose, the individual observation devices 11, 12, 13 are pulse-coupled oscillator devices 21, 22, 23

assigned, which possess the property, the automatic Synchronization to accomplish. The pulse-coupled oscillator devices 21, 22, 23 can be designed as software or hardware.

Before discussing this application, the mode of operation of the pulse-coupled oscillator devices 21, 22, 23 will be discussed.

In nature are automatically synchronizing

Phenomena are known (see, e.g., Mirollo, Strogatz;

Synchronization of pulse-coupled biological oscillators; SIAM J. Appl. Math, Vol. 50, No. 6, pp. 1645-1662, December 1990; Ermentrout et al. ; Beyond a Pacemaker's entrainment limit: Phase walk through. At the. J. Physiol., 1984 Jan, 246 (1 Pt 2): R 102-106).

Characteristic of these systems is that distributed, similar and related units initially exhibit different dynamic behavior (e.g., emission of firefly beacons). A very simple signal exchange between the fireflies finally causes after a certain time automatically a

Setting synchronization; all the fireflies that are in contact with each other flash synchronously.

By way of example, such an automatic synchronization of oscillators is described in the publications cited above. The described pulse-coupled

Oscillator devices 21, 22, 23 have a charging function v (t) (charging function, based on the charging process of a capacitor). In Fig. 2a is such

Charging function v (t) shown as an example.

The charge function v (t) is concave (v '' (t) <0), strictly monotone increasing (v '(t)> 0) and smooth in normalized

Interval [0, 1]. The state function v (y-axis) is also normalized to an interval [0, 1]. In the following, however, the charging function v (t) is used as a function of the time t.

Such an oscillator device 21, 22, 23 undergoes a continuous cycle of charging and discharging phases; it lies an oscillation before. The charging function v (t) increases from v (0) = 0 until the threshold v _max = 1 is reached. Then the charge function v (t) is reset to 0; in a figurative sense, a discharge has taken place.

From the period from t = 0 until reaching the

Thresholds v _Max can be the firing frequency of the

Derive oscillator device 21, 22, 23; the shorter the time interval, the higher the frequency of the

Oscillator device 21, 22, 23.

When the oscillator device 21, 22, 23 has reached the threshold value v _Max , it sends out a signal 30. In nature this can be a noise; in data processing this is usually an electrical signal or a

Data packet 30. However, an acoustic signal 30, an optical signal and / or a mechanical signal (for example pulse) 30 may also be used. In principle, combinations of signals 30 of different types are possible.

On the basis of nature, the emission of the signal 30 after reaching the threshold value V _{Max is} also referred to as "firing", the oscillator device 21, 22, 23 has a firing frequency.

The more frequently the threshold value v _{Max is} reached, the more frequently a signal 30 is sent.

Due to the monotonically increasing function v (t), a reduction of the threshold value v _Max causes the time until the transmission of the signal 30 to be shortened. This means that the "firing frequency" of the oscillator device 21, 22, 23 increases.

The oscillator devices 21, 22, 23 are designed not only to emit but also to receive signals 30. The receipt of a signal 30 at an oscillator device 21, 22, 23 causes the

State variable v of the oscillator device 21, 22, 23 is raised by an amount ε, as shown first in FIG. 2B. It has been shown mathematically, inter alia, in the publications cited above, that identically constructed

Pulse-coupled oscillator devices 21, 22, 23 with these characteristics that can exchange signals 30 with each other automatically synchronize after a certain time. Each oscillator device 21, 22, 23 simply follows the given rules; the phase synchronization of all oscillator devices 21, 22, 23 then adjusts automatically.

It is not necessary to have a central computer for

Checking or coordinating the measurements, since the synchronization automatically emerges as emergent

Property yields. Once the synchronization is achieved, it is stable.

Applied to the observation devices 11, 12, 13 in data transmission networks 100, this means that even with temporal changes in the data volume, which is measured by the correspondingly controlled observation devices 11, 12, 13, automatically recurs again and again

synchronous behavior sets.

This principle is shown in the context of

Embodiments on the automatic synchronization of the observation devices 11, 12, 13 in one

Data transmission network 100 is used. The synchronization of the observation devices 11, 12, 13 is somewhat different than e.g. the synchronization of timers in a network, as e.g. from Hong et al. , IEEE Journal on Selected Areas in Communications, pp. 1085-1099, May 2005. It was about the influence of

Time delays in data transmission.

In an advantageous embodiment, the above

described method of loading and unloading the

Charging function v (t) of the pulse-coupled oscillator device 21, 22, 23 modified. It is from the current

FeuerTeitpunkt to, a selection probability p _s for the observation devices 11, 12, 13 generated. The determination of this selection probability p _s will be explained below in connection with FIG. 3.

In FIG. 3, in the time domain, the charging function v (t)

is shown, which charges in a normalized time interval [0, 1] from v = 0 to v = l. The discharge time t = l corresponds to a maximum value for the charging time

Conversely, the maximum recharge time corresponds to one

minimum fire frequency. If the oscillator device 21, 22, 23 takes a long time to recharge, then it fires

less common .

If the threshold of v _Ma x = l to a new

Threshold v 'Max 1 is reduced, then this leads to the associated new fire time t'o occurring earlier; this means that lowering the

Threshold v _Max leads to a faster rate of fire.

However, the threshold v _max can not be arbitrary

be lowered, since the oscillator device 21, 22, 23 would otherwise fire infinitely fast. There is thus a v _M i _n that the smallest possible firing time (t _M i _n )

corresponds and thus the highest possible fire frequency.

Now a connection between the new

Fire times t '(t _M in <t'<t _max ), the new

Threshold values (v _M i _n <v ' _Ma x <v _max ) and the actual selection probabilities p' _s (p _M i _n <P 's <PMax) for the observation devices 21, 22, 23. In most cases it makes sense to set the limits p _M i _N = 0 (no package is selected) and p _max = 1 (all packages are selected). However, Ρ _ΜΙΠ and PMax can be set as desired, but p _MIN > 0 and p _max -1 are always valid ·

It is assumed that between the

Threshold v _Max and the selection probability p _{s is} a positive correlation. This means that a relatively small threshold v _Max leads to a relatively small selection probability p _s . A relatively large one

Threshold v _Max leads to a relatively high Selection probability p _s . There is a

deterministic coupling between the threshold value Vj4 _ax and the selection probability p _s . At regular intervals, the selection probability p _{s is} calculated or determined based on the current oscillator frequency. If a monitoring point has to lower the selection probability due to high CPU utilization, for example, it will change the

Threshold v _{Max of} its associated oscillator device 21, 22, 23.

In Fig. 3 is a second x-axis with the

Selection probability p _s drawn.

There is a negative correlation between the selection probability p _s and the frequency of the oscillator device 21, 22, 23, because the higher the frequency of the

Oscillator device 21, 22, 23, the lower the selection probability p _s .

The following is a functional relationship between the new threshold v 'ax and the new one

Selection probability p ' _s described.

The new threshold v ' _Max can only be between v _max and v _min = 1- ε / a with a> 0, eg a = 2.

As can be seen in FIG. 3, the minimum is

Selection probability p _M i _n the minimum threshold v _M i _n = l - ε / a with a> 0, for example a = 2 assigned. The maximum selection probability p _Ma x is assigned the maximum threshold v _max = 1.

If the relationship between the new threshold v ' _Max and the new selection probability p' _s , given by v (t), is assumed to be linear (in FIG

dashed line drawn), one can establish the following equation of proportion on the basis of the track conditions on the x-axis for p and the y-axis for v:

a In the case where v 'Max ~ (ie the threshold has the theoretically maximum value), the maximum

Selection probability p ' _s = p _max reached.

In the event that v'Max-1 ^- ε / a - i4 ^ _n , the minimum selection probability p ' _s = p _M i _{n is} reached.

In both cases, of course, the reverse apply

Statement .

From equation (1) a new selection probability p ' _s can be calculated for each new threshold v' _Max and vice versa.

In other embodiments, the functional one is

Relationship between v 'ax and p _s non-1inear, so that the calculation rules are more complex, but basically work on the same principle. In any case, the equation (1) represents an assignment function (mapping function), with which from the changing

Threshold values v 'ax and the concomitant changes in the firing times t finally new

Selection probability p ' _s can be found.

As a result, the original value range of the

State variables v (see Fig. 2) are automatically normalized to a new range, so that the new threshold v ' _{Max is} smaller and thus the pulse-coupled

Oscillator device 21, 22, 23 "discharges" faster, ie emits signals 30 with a higher frequency, which then leads to a lower selection probability p ' _s due to the assignment function made.

That of the highest firing frequency oscillator devices 21, 22, 23 dominates all others with which it is associated. By the transmitted signals 30 is the selection probability of all others

Oscillator devices 21, 22, 23 to the new

Selection probability p ' _s changed. By the above conversion rule can be any

Observation device 11, 12, 13 based on their associated oscillator device 21, 22, 23, the new Determine selection probability p ' _s . That is to say, all observation devices 11, 12, 13 run on the lowest selection probability p _s after this process. The principle is: the highest frequency of

Oscillator device 21, 22, 23 (ie the smallest

FeuerTeitpunkt to) corresponds to the lowest new selection probability p ' _s -

In connection with FIG. 4 it is shown how the pulse coupled oscillator devices 21, 22, 23 can be used to synchronize the automatic checking of the data traffic.

In this case, it is assumed, for example, of three observation devices 11, 12, 13 at which data for monitoring the data flow in the data transmission network 100 are collected. In this case, the new selection probability p ' _{s of} the observation devices 11, 12, 13 in the above

be associated with the frequency of the pulse-coupled oscillator devices 21, 22, 23 described manner. Each observation device 11, 12, 13 is for this purpose at least with such a pulse-coupled

Oscillator device 21, 22, 23 coupled. In principle, it is also possible that a pulse-coupled

Oscillator device 21, 22, 23 with more than one

Observation device 11, 12, 13 is coupled.

In the case that an oscillator device 21 their

Threshold v = v reaches _Max , it sends an IP packet as a signal 30 to all other oscillator devices 22, 23, the respective observation devices 11, 12, 13th

assigned. These receive the signal 30 and

process it as described in connection with Fig. 2, i. the value of the charging function v (t) is increased by the amount ε, so that this

Oscillator devices 22, 23 are brought closer to the discharge.

Such a total system has the property of synchronizing itself, ie the selection probability p _{s of} the observation devices 11, 12, 13 become automatic synchronized by the characteristics of the associated pulse-coupled oscillator devices 21, 22, 23. It always dominates the highest frequency, ie amount of

Oscillator devices 21, 22, 23 adapts to the highest frequency. Due to the assignment made above, this highest frequency is assigned the lowest selection probability p ' _s .

As shown in Fig. 4, after some time in all observation devices 21, 22, 23, the new

Selection probability p ' _s before.

Thus, each observation device 11, 12, 13 capable of the given functional relationship (e.g., equation (1)) is the required new one

Selection probability p ' _s to calculate.

In this way it is possible that throughout

Data transmission network 100, the selection probability p _{s is} automatically lowered, if the data flow increases too much. A central monitoring is not necessary because the automatically synchronizing properties of the

pulse coupled oscillator devices 21, 22, 23

be exploited.

Reference sign list

11, 12, 13 observation device

21, 22, 23 pulse coupled oscillator device

30 signal, data packet

100 Data transmission network p _s Selection probability of

Observer p _'s new selection probability or

Observation device to firing point of time v (t) Charging function v _Max upper threshold for charging function v ' _Max new upper threshold for charging function v _Min lower threshold for charging function

Claims

A method for synchronizing multipoint measurements in a data transmission network (100) having at least two observation devices (11, 12, 13) performing measurements, a) wherein the at least two observation devices (11, 12, 13) have at least two pulse-coupled ones

Oscillator devices (21, 22, 23) are assigned, b) the at least two pulse-coupled

Oscillator devices (21, 22, 23) one each

Have a charge function v (t) that ranges from v (t = 0) = 0 to

v (t = t _max ) = v _Max increases and each of the pulse-coupled ones

Oscillator devices (21, 22, 23) when reaching the

Threshold value v _Max automatically sends a signal (30) to other pulse coupled oscillator devices (21, 22, 23) and then set to v = 0, the time interval 0 to t _{Max determining} the frequencies of the oscillator devices (21, 22, 23), c) in which the signal (30) receiving pulse-coupled oscillator devices (21, 22, 23) the charging function v (t) is raised by the amount ε, d) so that all together by signals (30) coupled oscillator devices (21, 22, 23) and thus also the coupled observation devices (11, 12, 13) due to the self-synchronizing properties of

pulse-coupled oscillator devices (21, 22, 23) perform synchronous measurements in the data transmission network, wherein e) a synchronized selection probability p _{s of} the observation devices (11, 12, 13) for the traffic from the functional relationship between the loading function v (t) and the selection probability p _s is formed so that the selection probability p _s negative with the

Frequency of the oscillator device (21, 22, 23) correlated.

2. The method according to claim 1, characterized

characterized in that the functional relationship between the loading function v (t) and the selection probability is formed such that the new threshold 't ax correlates positively with the new selection probability p' _s .

3. The method according to claim 1 or 2, characterized

characterized in that the charging function v (t) is strictly monotonically increasing and concave.

4. The method according to at least one of the preceding

Claims, characterized in that the

Charging function v (t) of the oscillator devices is at least partially linearized to a relationship between a new threshold v'Max and a new

Selection probability p _s to find.

5. The method according to claim 4, characterized

characterized in that the relationship between the new threshold v 'ax and the new

Selection probability p _s through

a

(for a> 0) is given.

6. Method according to at least one of the preceding

Claims, characterized in that the at least one observation device (11, 12, 13) is adaptive to a change in the data transmission in the

Data transmission network (100) responds.

7. The method according to at least one of the preceding

Claims, characterized in that all pulse-coupled oscillator devices (21, 22, 23) of the observation devices (11, 12, 13) have substantially the same dynamics and in the data transmission network (100) are coupled together, in particular exchange signals (30).

8. The method according to at least one of the preceding

Claims, characterized in that the

Synchronization of the selection process using a hash function is done.

9. The method according to claim 9, characterized

characterized in that the selection process of packages is done by means of a hash selection area.

10. Device for synchronizing measurements

Data packets in a data transmission network (100) with at least two observation devices (11, 12, 13) for performing measurements, a) wherein the at least two observation devices (11, 12, 13) at least two pulse-coupled

Oscillator devices (21, 22, 23) are assigned, b) the at least two pulse-coupled

Oscillator devices (21, 22, 23) one each

Have a charge function v (t) that ranges from v (t = 0) = 0 to

v (t = t _M ax) = v x _Ma increases and each of the pulse-coupled

Oscillator devices (21, 22, 23) when reaching the

Threshold v _Max automatically sends a signal (30) to other pulse coupled oscillator devices (21, 22, 23) and then set to v = 0, the time interval 0 to t _Ma x determining the frequencies of the oscillator devices (21, 22, 23) c) in which the pulse-coupled oscillator devices (21, 22, 23) receiving the signal (30) increase the charging function v (t) by the amount ε, d) a synchronized selection probability p _{s of} the observation devices (11, 12, 13) for the data traffic is formed from the functional relationship between the charging function v (t) and the selection probability such that the selection probability p _{s is} negative with the frequency of the oscillator device (21 , 22, 23).

11. The device according to claim 10, characterized

characterized in that the functional relationship between the charging function v (t) and the

Selection probability is formed so that the new threshold v'Max positive with the new

Selection probability p ' _s correlates.

12. The device according to claim 10 or 11, characterized in that the charging function v (t) is strictly monotonically rising and concave.

13. The device according to at least one of claims 10 to 12, characterized in that the charging function v (t) of the oscillator devices (21, 22, 23) is at least partially linearized to a relationship between a new threshold v 'ax and a new

Selection probability p _s to find.

14. The apparatus according to claim 13, characterized

Selection probability p _s through

a is given.

15. The device according to at least one of claims 10 to 14, characterized in that all

pulse-coupled oscillator devices (21, 22, 23) of the observation devices (11, 12, 13) have substantially the same dynamics and in the data transmission network (100) are coupled together, and in particular exchange signals (30).