GB2560515A - Telecommunications network - Google Patents
Telecommunications network Download PDFInfo
- Publication number
- GB2560515A GB2560515A GB1703942.1A GB201703942A GB2560515A GB 2560515 A GB2560515 A GB 2560515A GB 201703942 A GB201703942 A GB 201703942A GB 2560515 A GB2560515 A GB 2560515A
- Authority
- GB
- United Kingdom
- Prior art keywords
- traffic
- sets
- peak
- observations
- link
- 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.)
- Granted
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
- H04L41/147—Network analysis or design for predicting network behaviour
-
- 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
- H04L43/0876—Network utilisation, e.g. volume of load or congestion level
- H04L43/0894—Packet rate
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/50—Network service management, e.g. ensuring proper service fulfilment according to agreements
- H04L41/5003—Managing SLA; Interaction between SLA and QoS
- H04L41/5019—Ensuring fulfilment of SLA
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
A method to estimate the required capacity of a network link. A first set of traffic observations is made by counting packets within each interval with a duration Y that is contained within time periods N1-N3 of a peak utilisation period. A second set of observations is made for intervals of duration X. Peak and average data rates are determined for each set of traffic observations within each time period N1-N3. A peak to average data rate ratio is calculated for each observation interval set in each time period N1-N3. One of the sets is selected for each observation interval from one of the time periods N1-N3 e.g. the one with a maximum peak to mean value. The selected peak to average ratios are then interpolated against the observation interval and a goodness to fit value determined for a power function. This value relates to the degree of burstiness of the traffic and is used to select which observation interval period should be used for dimensioning the link.
Description
(54) Title of the Invention: Telecommunications network
Abstract Title: Determining a peak to average data rate ratio from a plurality of traffic observation sets to prevent over-estimation of bandwidth for a network link (57) A method to estimate the required capacity of a network link. A first set of traffic observations is made by counting packets within each interval with a duration Y that is contained within time periods N1-N3 of a peak utilisation period. A second set of observations is made for intervals of duration X. Peak and average data rates are determined for each set of traffic observations within each time period N1-N3. A peak to average data rate ratio is calculated for each observation interval set in each time period N1-N3. One of the sets is selected for each observation interval from one of the time periods N1-N3 e.g. the one with a maximum peak to mean value. The selected peak to average ratios are then interpolated against the observation interval and a goodness to fit value determined for a power function. This value relates to the “degree of burstiness” of the traffic and is used to select which observation interval period should be used for dimensioning the link.
-Peak Utilisation Period-
<-r Time Period, N,-► X X X X X X | <-2 Time Period, N2-► X X X X X X | |
mil III | ϊίΐϊΐ II II |
yyyyy yyyyy „ Figure 5
X = 2 Observation Interval
Y= 1st Observation Interval
-3rd Time Period, N3-►
X X X X X X mil I I I I
Time
1/6
6θ3ν
Figure 1
2/6
Figure 2
3/6
Figure 3
4/6
(51) (52) (53) (54) (55) (56) (57) (58)
Figure 4
5/6
CD
E
Figure 5
I
I
I
6/6
Figure 6
Intellectual
Property
Office
Application No. GB1703942.1
RTM
Date :23 August 2017
The following terms are registered trade marks and should be read as such wherever they occur in this document:
GFI LanGuard (page 13)
Icinga (page 14)
Microsoft (page 13)
Nagios (page 13)
Pandora FMS (page 13)
Splunk (page 14)
Wireshark (page 13)
Intellectual Property Office is an operating name of the Patent Office www.gov.uk/ipo
TELECOMMUNICATIONS NETWORK
Field of the Invention
The present invention relates to a method of dimensioning a link in a telecommunications network.
Background to the Invention
Telecommunications networks include a plurality of network nodes connected by links (such as optical fibre connections). When planning or modifying a network, the Network Operator must estimate the capacity of a link between two network nodes (typically known as “dimensioning”) to accommodate the expected traffic on that link. This is particularly important when the Network Operator must meet certain Service Level Agreement (SLA) targets.
In most networking scenarios, a link carries data for a plurality of users and there is usually a direct relationship between the number of users for which a link carries traffic for and the smoothness of a traffic profile for the link. That is, as the number of users of a link increases, the traffic profile for that link is relatively smooth. Accordingly, links between core networking nodes have relatively smooth traffic profiles compared to links between edge networking nodes or those in enterprise networks. In this sense, a relatively smooth traffic profile is one having fewer and/or less pronounced spikes from a mean traffic rate (in which other variables, such as time of day, remain constant).
Network Operators may operate links at high utilisations (e.g. average traffic at 80% or more of the link capacity) when it is known the traffic is relatively smooth. Nonetheless, Operators typically upgrade links when their utilisation reaches certain thresholds, such as 40-50% of link capacity. However, when the traffic profile is less smooth due to the presence of traffic bursts (that is, in which the instantaneous traffic produces one or more pronounced spikes above the mean traffic rate) then a highly utilised link may suddenly be overwhelmed. This may result in packets being queued at network nodes associated with the link, resulting in packet delays, discards and jitter. Such performance degradation may even be realised on links which are operating at a very low utilisation if, for example, a traffic burst is sufficiently demanding and/or the link utilisation measurements of peak and mean traffic rates are taken over coarse timescales.
A well-known practice for link dimensioning is to multiply the mean traffic rate by a “peakto-mean” factor to estimate the capacity of a link. Typically, Network Operators use a value of 30 for the peak-to-mean factor to accommodate the expected traffic on a link and meet the SLA requirements. However, in some scenarios, this value is overly generous and results in greater than necessary costs for building a network to satisfy the SLA requirements (as higher-capacity links tend to cost more than lower-capacity links).
It is therefore desirable to alleviate some or all of the above problems.
Summary of the Invention
According to a first aspect of the invention, there is provided a method of dimensioning a link in a telecommunications network, the method comprising the steps of: determining, for a first plurality of sets of traffic observations, wherein each set of the first plurality of sets includes a series of traffic observations each indicating the amount of traffic on a link within an observation interval of a particular duration for that set, a peak value of a traffic observation of the series of traffic observations within each set of the first plurality of sets and an average value of the series of traffic observations within each set of the first plurality of sets; calculating a first peak to average ratio for each observation interval duration based on the determined peak and average values within each set of the first plurality of sets; determining a goodness of fit value for a power function of the first peak to average ratio against its respective observation interval duration; selecting a dimensioning interval duration based on the determined goodness of fit value; and estimating a capacity of the link based on the selected dimensioning interval duration.
According to a second aspect of the invention, there is provided a computer program comprising instructions which, when the program is executed by a computer, cause the computer to carry out the method of the first aspect of the invention. The computer program may be stored on a computer readable data carrier.
According to a third aspect of the invention, there is provided a device having a processor for performing the method of the first aspect of the invention.
The step of selecting the dimensioning interval duration may be based on the goodness of fit value and a service level requirement.
The first plurality of sets of traffic observations are within a first time period, and the method may further comprise the steps of: determining, for a second plurality of sets of traffic observations within a second time period, wherein each set of the second plurality of sets includes a series of traffic observations each indicating the amount of traffic on the link within an observation interval of a particular duration for that set, a peak value of a traffic observation of the series of traffic observations within each set of the second plurality of sets and an average value of the series of traffic observations within each set of the second plurality of sets, wherein the observation interval durations for each set of the first plurality of sets of traffic observations are equal to the observation interval durations for each set of the second plurality of sets of traffic observations.
The method may further comprise the step of calculating a second peak to average ratio for each observation interval duration based on the determined peak and average values within each set of the second plurality of sets, wherein the power function may be of either the first or second peak to average ratio against its respective observation interval duration.
The power function may be of the maximum of either the first or second peak to average ratio against its respective observation interval duration.
The method may further comprise the step of: determining a peak-utilisation period for the link, wherein the first and second time periods are within the peak-utilisation period.
The method may further comprise the steps of: determining first and second overall traffic observations for the first and second time period respectively, wherein the first and second overall traffic observations indicate the amount of traffic on the link in the first and second time periods respectively; and filtering the series of traffic observations of the first and second plurality of sets of any traffic observation that satisfies a criterion based on the first and second overall traffic observations respectively.
One of the series of observations of a first set of the first and/or second plurality of sets may occur at the same time instance as one of the series of observations of a second set of the first and/or second plurality of sets.
Brief Description of the Figures
In order that the present invention may be better understood, embodiments thereof will now be described, by way of example only, with reference to the accompanying drawings in which:
Figure 1 is a schematic diagram of an embodiment of a telecommunications network of the present invention;
Figure 2 is a schematic diagram of a Digital Subscriber Line Access Multiplexer (DSLAM) of Figure 1;
Figure 3 is a schematic diagram of a Network Management System (NMS) of Figure 1; Figure 4 is a flow diagram of an embodiment of a method of the present invention; Figure 5 is a diagram illustrating a peak-utilisation hour of a link of Figure 1; and Figure 6 is a graph illustrating several peak-to-mean ratios against observation interval of the link of Figure 1.
Detailed Description of Embodiments
A first embodiment of a telecommunications network 1 of the present invention will now be described with reference to Figures 1 to 3. The telecommunications network 1 includes a plurality of network nodes, including a plurality of Customer Premises Equipment (CPE) 10a... 10h, a first and second Digital Subscriber Line Access Multiplexer (DSLAM) 20, 30, an access network gateway router 40 and a core network router 50. The first and second DSLAMS 20, 30, access network gateway router 40 and core network router 50 are each connected to a respective Network Management System (NMS) 60a, 60b, 60c, 60d.
Figure 1 also illustrates a plurality of links, including a first link 70 between a first CPE 10a and the first DSLAM 20, a second link 80 between the first DSLAM 20 and the access network gateway router 40, and a third link 90 between the access network gateway router 40 and the core network router 50. The first link 70 carries traffic between the first CPE 10a and the first DSLAM 20 only. The first DSLAM 20 multiplexes this traffic with any traffic associated with the respective links with the second, third and fourth CPEs 10b, 10c, 10d, which is thereafter carried on the second link 80 to the access network gateway router 40. The access network gateway router 40 multiplexes this traffic with any traffic associated with the respective link with the second DSLAM 30 (and therefore the fifth, sixth, seventh and eighth CPE 10e, 10f, 10g, 10h), which is thereafter carried on the third link 90 to the core network router 50. This process is often known as “link aggregation”. Accordingly, the third link 90 requires more capacity than the second link 80, and the second link 80 requires more capacity than the first link 70.
The first DSLAM 20 is shown in more detail in Figure 2. The first DSLAM 20 includes a first transceiver 21, a processor 23, memory 25, and a second transceiver 27, all connected via bus 29. The first transceiver 21 is an optical fibre interface for a downstream connection to the first, second, third and fourth CPEs 10a, 10b, 10c, 10d (wherein the connection to the first CPE 10a is via the first link 70). The second transceiver 27 is also an optical fibre interface for an upstream connection to the access network gateway router 40 via the second link 80, and a connection to the NMS 60a. The processor 23 includes a packet counting function which is adapted to count the number of packets transmitted to any one of the CPEs 10a, 10b, 10c, 10d via the first transceiver 21 or to the access network gateway router 40 via the second transceiver
27. In this embodiment, the packet counting function is implemented using traffic sniffing software, Wireshark ®, provided by Wireshark Foundation.
In one particular use of the packet counting function, the processor 23 is configured to count the number of packets in five and fifteen minute time periods transmitted over the first link 70 to the first CPE 10a, and store a value for the average (mean) data rate in these five and fifteen minute periods in memory 25. The processor 23 is also configured to count the number of packets in five and fifteen minute time periods transmitted over the second link 80 to the access network gateway router 40, and store a value for the average data rate in these five and fifteen minute periods in memory 25. Memory 25 stores a plurality of these values for these five and fifteen minute average data rates in a Management Information Base (MIB), including the most recent and a plurality of historical values (up to, for example, a day’s worth). The first DSLAM 20 is configured to report these values for the five and fifteen minute average data rates for the first and second links 70, 80 to the NMS 60a (either periodically or in response to a request).
In this embodiment, the plurality of CPEs 10a... 10h, the second DSLAM 30, the access network gateway router 40 and the core network router 50 also contain first and second transceivers, processors and memory, wherein the processors contain packet counting functions for counting the number of packets transmitted via the first and/or second transceivers.
The NMS 60a is shown in more detail in Figure 3. The NMS 60a includes a first transceiver 61, a processor 63 and memory 65, all connected via bus 67. The first transceiver 61 is an optical fibre interface for connecting the NMS 60a to the first DSLAM 20. Memory 65 is configured to store a plurality of values of the five and fifteen minute average data rates (that is, the most recent values and the historical values) for the first and second links 70, 80.
In this embodiment, the NMSs associated with the second DSLAM 30, the access network gateway router 40 and the core network router 50 are substantially the same. However, the skilled person will understand that it is possible for a single NMS to be associated with multiple nodes in the network.
An embodiment of a method of the present invention will now be described with reference to Figures 4 to 6. This embodiment relates to the dimensioning of the second link 80 (between the first DSLAM 20 and the access network gateway router 40) in an operational network, such that the Network Operator is re-provisioning (or upgrading) the second link 80.
As a first step (step S1), NMS 60c retrieves the plurality of values of the five and fifteen minute average data rates (stored in memory 65) to profile the average data rate over the day and determine a peak-utilisation period for the second link 80 (e.g. the hour in the day having the greatest average data rate). The NMS 60c then identifies Λ/ disjoint time periods within the peak-utilisation period, and obtains average data rates for these /V time periods. In this embodiment, these Λ/ time periods and their average data rates are three of the five minute average data rates for the second link 80 separated by fifteen minutes within the peak-utilisation period.
In the second step (step S2), NMS 60c instructs the packet counting function of the processor of the access network gateway router 40 to count the number of packets transmitted over the second link 80 to obtain a series of traffic measurements within each of the Λ/ time-periods. In this embodiment, the access network gateway router 40 counts the number of packets transmitted within successive observation intervals of a particular duration (such as, for example, a millisecond), such that it obtains a count of packets in the first millisecond, a count of packets in the second millisecond, and so on, for the entirety of each of the Λ/ time-periods. This first set of traffic observations is stored in memory.
The access network gateway router 40 reports the first set of traffic observations (each indicating the count of packets within successive observation intervals of a particular duration) to the NMS 60c, which are stored in memory 65. The processor 63 is then able to evaluate further sets of traffic observations for the Λ/ time periods, wherein each set indicates a count of packets transmitted within successive observations intervals, wherein the observation interval duration for each set is different. For example, a second, third, fourth and fifth set of traffic observations relate to observation interval durations of ten milliseconds, fifty milliseconds, one-hundred milliseconds, and onesecond respectively, and may be evaluated by summing the corresponding counts from the first set of traffic observations indicating the count of packets within successive onemillisecond observation intervals. Thus, the second set of traffic observations indicating the count of packets in successive ten-millisecond intervals may be evaluated by summing the counts of packets in the traffic observations for the first to the tenth onemillisecond observation intervals, the counts of packets in the traffic observations for the eleventh to the twentieth one-millisecond observation intervals, and so on. This processing is performed to evaluate a plurality of sets of traffic observations (each set relating to a different observation interval duration) for each of the three time periods to create a first, second and third plurality of sets of traffic observations.
Accordingly, once this processing is complete, the NMS 60a stores a first plurality of sets of traffic observations, including the first, second, third, fourth and fifth sets of traffic observations for the first time period (that is, the first set of traffic observations indicating the count of packets at a first observation interval duration in the first time period, the second set of traffic observations indicating the count of packets at a second observation interval duration in the first time period, and so on), a second plurality of sets of traffic observations, including the first, second, third, fourth and fifth sets of traffic observations for the second time period (that is, the first set of traffic observations indicating the count of packets at the first observation interval duration in the second time period, the second set of traffic observations indicating the count of packets at the second observation interval duration in the second time period, and so on), and a third plurality of sets of traffic observations, including the first, second, third, fourth and fifth sets of traffic observations for the third time period (that is, the first set of traffic observations indicating the count of packets at the first observation interval duration in the third time period, the second set of traffic observations indicating the count of packets at the second observation interval duration in the third time period, and so on).
For completeness, the concept and relationships between the peak-utilisation period, the disjoint time periods, and the first, second and third plurality of sets of traffic observations of differing interval durations are illustrated in Figure 5. Firstly, it is noted that the term “time period” represents elapsed time having a defined start and end point, and the term “interval” represents elapsed time between any start and end point. In Figure 5, the horizontal axis represents time, and the diagram focusses on a peak-utilisation time period. Figure 5 also illustrates the 1st, 2nd and 3rd disjoint time periods, Ni, N2, N3, which have defined start and end points and are non-overlapping. As noted above, the access network gateway router 40 counts the number of packets transmitted within successive observation intervals within the whole of each of the Λ/ time periods. As shown in Figure 5, these traffic observations relate to the number of packets that are transmitted in an observation interval (of duration y), and these are counted in successive observation intervals for the whole of each of the Λ/ time periods (although only the first five observation intervals of each of the Λ/ time periods are shown in Figure 5). This data, indicating each successive traffic observation (of length y) and the count of packets in each traffic observation, is then reported to the NMS 60c as the first set of traffic observations.
The NMS 60c then evaluates a second set of traffic observations, wherein the second set of traffic observations uses observation interval duration ‘x’, within the whole of each of the Λ/ time periods. This is achieved by summing the counts of packets from the first set of traffic observations of observation interval duration ‘y’. For example, the first traffic observation of the second set (of duration ‘x’) is the sum of all of the first five traffic observations of the first set (of duration ‘y’). Although Figure 5 is not to scale and only shows two sets of traffic observations, it will now be clear how these concepts may be extended when each time period is subdivided into any number of sets of particular observation interval durations.
In an example, the NMS 60c has the following data:
First Set of Traffic Observations | Second Set of Traffic Observations | Third Set of Traffic Observations |
(Interval Duration: 1ms) | (Interval Duration: 10ms) | (Interval Duration: 50ms) | |||||
Time Period (W) | Interval (ms) | Count of Packets | Interval (ms) | Count of Packets | Interval (ms) | Count of Packets | |
/Vi | 0->1 | 58 | 0->10 | 475 | 0->50 | 2541 | |
1->2 | 69 | 10->20 | 577 | 50->100 | 2307 | ||
2->3 | 0 | 20->30 | 496 | 100->150 | 2466 | ||
3->4 | 60 | 30->40 | 479 | 150->200 | 2137 | ||
4->5 | 82 | 40->50 | 514 | 200->250 | 2412 | ||
N2 | 0->1 | 1 | 0->10 | 437 | 0->50 | 2434 | |
1->2 | 67 | 10->20 | 461 | 50->100 | 2793 | ||
2->3 | 75 | 20->30 | 472 | 100->150 | 2578 | ||
3->4 | 46 | 30->40 | 479 | 150->200 | 2563 | ||
4->5 | 17 | 40->50 | 458 | 200->250 | 2636 | ||
n3 |
Table 1: Τable illustrating a count of packets within a first, second and third sets of traffic observations, stored in memory 65
The NMS 60c may then translate these packet count values into data rates by multiplying the packet count by the packet size, and dividing by the respective observation interval duration. For example, the data rate for 40->50ms of the second set in the first time period, Ni, would be (514 packets * 12000 bits/packet)/10ms, which equals 616800000 bits/s (or 616.8Mb/s).
In this embodiment, NMS 60c is configured to filter these results in order to remove any result which is not within a threshold of the five minute average data rate of the relevant time period (determined in step S1). This is achieved by setting a threshold parameter, δ, and filtering any result which is not within the five minute average data rate (for the relevant time period) plus or minus δ. For example, if δ is 0.01 and the five minute average data rate is 615MB/S, then any result which is not within 615*0.99=608.85 and
615*1.01=621.15 is filtered out. This has the benefit of removing any outliers from the results.
In step S3, the NMS 60c determines the peak filtered data rate of each set of traffic observations for each of the Λ/ time periods, and the mean filtered data rate within each set of traffic observations for each of the N time periods. In step S4, the NMS 60c selects a subset of the sets of traffic observations, being those which have a mean filtered data rate being substantially the same (e.g. wherein the mean filtered data rate for each of the subset of set of traffic observations is within a threshold, δ). In this example, the NMS 60c selects the first, second, fourth and fifth sets of traffic observations (that is, having the one-millisecond, the ten millisecond, the one-hundred millisecond and the one-second observation interval durations). Again, this has the benefit of removing outlier results from the subsequent processing steps.
The NMS 60c then determines a peak to mean ratio for each observation interval duration. This determination is made in two steps. Firstly, the NMS 60c calculates a peak to mean ratio for each of the subset of the sets of traffic observations for each time period, based on the peak filtered data rate and mean filtered data rate for each set of traffic observations. This data is illustrated in the following table:
Time Period (W) | First Set of Traffic Observations | Second Set of Traffic Observations (Interval Duration: 10ms) | Fourth Set of Traffic Observations (Interval Duration: 100ms) | |
(Interval 1ms) | Duration: | |||
/Vi | Ψ(ΐΓηδ,ΝΙ) | Ψ(ΙΟπδ,ΝΙ) | Ψ(ΙΟΟπδ,ΝΙ) | |
N2 | Ψ(1γπ5,Ν2) | Ψ(10π5,Ν2) | Ψ(100π5,Ν2) | |
n3 | Ψ(1πδ,Ν3) | Ψ(10πδ,Ν3) | Ψ(100πδ,Ν3) | |
Table 2: Table il | ustrating the peak to mean ratios calcu | ated for each set of traffic |
observations for each time period
In step S5, the NMS 60c selects one of the peak to mean ratios for each set of traffic observations from the three time periods (e.g. one of Ψ(ΐπδ,Νΐ), Ψ(ΐπΐ5,Ν2) or Ψ(ΐηΐ5,Ν3) for the first set of traffic observations). In this embodiment, the NMS 60c selects the maximum peak to mean ratio for each set of traffic observations across the three time periods (for example, the selected peak to mean ratio for the first series Ψ(ΐηΊδ) is the maximum of Ψ(ΐπδ,Νΐ), Ψ(1πί5,ν2) and Ψ(ΐηΐ5,Ν3)). This maximum value is then the selected peak to mean ratio for the corresponding observation interval duration.
A theoretical discussion of the concepts behind the remaining steps will now be presented. In testing scenarios, the present inventors evaluated the selected peak to mean ratios for each observation interval duration by interpolating them against observation interval duration according to a power function (of format Ψ=3*χ'0, where Ψ is the selected peak to mean ratio, x is the corresponding observation interval duration (e.g. 1ms, 10ms, 100ms or 1s), and a and b are constants). An example graph illustrating this data for one testing scenario is illustrated in Figure 6.
The present inventors further discovered, by performing such interpolations for a plurality of test links, that the goodness of fit of the power curve is indicative of the burstiness of traffic in each link. That is, for relatively bursty traffic (or in other words, less smooth traffic), the goodness of fit of a power function is relatively high, whereas for relatively smooth traffic (or in other words, less bursty traffic), the goodness of fit of a power function is relatively low. The method of interpolating data according to a power function and of determining a goodness of fit (such as the R-squared method) will be apparent to a person skilled in the art.
Accordingly, in step S6, the NMS 60c determines a goodness of fit value of a power function for the selected peak to mean ratios interpolated against observation interval duration. This value, hereinafter the “Degree of Burstiness” or “DB, is stored in memory 65, together with a timestamp, to indicate the burstiness of traffic of the second link 80 at that time.
In step S7, the NMS 60c determines a dimensioning interval duration (being one of observation interval durations of the first, second, fourth and fifth sets, e.g. 1ms, 10ms, 100ms, 1s), based on predetermined fuzzy logic rules. In this embodiment, these rules are:
1. If DB is relatively high (e.g. greater than a first DB threshold), and the link’s SLA requirements are stringent (e.g. satisfying one or more Operator defined thresholds), then a relatively short time duration is selected as the dimensioning interval duration, e.g. 1ms;
2. If DB is relatively high (e.g. greater than the first DB threshold), and the link’s SLA requirements are not stringent (based on the Operator defined thresholds), then a medium time duration is selected as the dimensioning interval duration, e.g. 10ms or 100ms;
3. If DB is relatively low (e.g. less than the first DB threshold), and the link’s SLA requirements are stringent (e.g. satisfying one or more Operator defined thresholds), then a medium time duration is selected as the dimensioning interval duration, e.g. 10ms or 100ms.
4. If DB is relatively low (e.g. less than the first DB threshold), and the link’s SLA requirements are not stringent (based on the Operator defined thresholds), then a relatively long time duration is selected as the dimensioning interval duration, e.g. 1s.
In an example, the first DB threshold may be 0.8, and the Operator defined SLA threshold may be that delay is less than 50ms for a link to have ‘stringent’ SLA requirements. However, the skilled person will understand that such thresholds may be tailored depending on the services provided by the link or by trailing test scenarios.
In step S8, the NMS 60c estimates the capacity of the link by multiplying the selected peak to mean ratio (selected in step S5) associated with the dimensioning interval duration (identified in step S7) by the second link’s average data rate. In this embodiment, the average data rate is that of the five minute time period associated with the selected peak to mean ratio. According to this process, any link which has relatively bursty traffic will be provisioned based on a peak to mean ratio associated with a relatively short dimensioning interval duration (compared to links with relatively smooth traffic). The peak to mean ratios for these shorter dimensioning interval durations are more sensitive to traffic bursts and are therefore greater than the peak to mean ratios for longer dimensioning interval durations. Accordingly, the NMS 60c will estimate greater capacities for links with relatively bursty traffic. Furthermore, as the burstiness of the traffic is being determined objectively by measurements on the link, the capacity of the link is proportional to its burstiness (based on the use of appropriate Operator-defined threshold), such that the link will not be over-provisioned or under-provisioned for its scenario.
The value for the estimated capacity is stored in memory 65 together with a timestamp. The Network Operator may therefore query the NMS 60c to determine the value for the estimated capacity. The Operator may then upgrade (or downgrade) the link to one having a capacity substantially equal to the estimated capacity. The actual capacity may be slightly greater than the estimate for resilience and predicted traffic growth purposes.
In this embodiment, the process loops back to step S1 (via a delay timer) and a new estimate of the capacity of the link is determined based on updated traffic measurements. In this manner, the Network Operator may query the NMS 60c to determine the most recent value for the estimated capacity.
The skilled person will understand that the number of time periods, and the number of sets of traffic observations in each time period, in the above embodiment are merely examples. However, a plurality of time periods are used in order to evaluate more data points across the peak-utilisation period. Furthermore, five sets of traffic observations are used in order to evaluate a power function for peak-to-mean ratio against observation interval duration, although fewer sets (and therefore fewer observation interval durations) may be used to interpolate a power function.
Furthermore, the peak-to-mean ratio for each observation interval duration is selected as the maximum of the peak-to-mean ratio for each set for each time period. The skilled person will understand that this is non-essential, and any of the peak-to-mean ratios for each set may be chosen. However, the maximum of these ratios will subsequently estimate a greater capacity for the link and is therefore a more conservative value.
The skilled person will also understand that it is non-essential that the data rate results are filtered of those that are not within a threshold, and that a subset of the sets of traffic observations are used. However, this has the benefit of removing outliers from the subsequent processing steps.
In the above embodiments, the packet counting function is implemented using Wireshark, but the skilled person will understand that other software may be used, such as GFI LanGuard, Microsoft Network Monitor, Nagios, OpenNMS, Advanced IP Scanner, Capda Free, Fiddler, NetworkMiner, Pandora FMS, Zenoss Core, PRTG Network
Monitor Freeware, The Dude, Splunk, Angry IP Scanner, Icinga 2, Total Network Monitor, NetXMS, and WirelessNetView.
In the above embodiments, the packet counting function observes the traffic on a link 5 and reports this data to an NMS, and the NMS derives further observations based on this data. However, the skilled person will understand that this is not the only way of implementing the present invention. For example, the packet counting function may perform observations of the traffic on the link for a plurality of observation intervals, which are thereafter reported to the NMS.
The skilled person will understand that any combination of features is possible within the scope of the invention, as claimed.
Claims (11)
1. A method of dimensioning a link in a telecommunications network, the method comprising the steps of:
determining, for a first plurality of sets of traffic observations, wherein each set of the first plurality of sets includes a series of traffic observations each indicating the amount of traffic on a link within an observation interval of a particular duration for that set, a peak value of a traffic observation of the series of traffic observations within each set of the first plurality of sets and an average value of the series of traffic observations within each set of the first plurality of sets;
calculating a first peak to average ratio for each observation interval duration based on the determined peak and average values within each set of the first plurality of sets; determining a goodness of fit value for a power function of the first peak to average ratio against its respective observation interval duration;
selecting a dimensioning interval duration based on the determined goodness of fit value; and estimating a capacity of the link based on the selected dimensioning interval duration.
2. A method as claimed in Claim 1, wherein the step of selecting the dimensioning interval duration is based on the goodness of fit value and a service level requirement.
3. A method as claimed in either Claim 1 or Claim 2, wherein the first plurality of sets of traffic observations are within a first time period, and the method further comprises the steps of:
determining, for a second plurality of sets of traffic observations within a second time period, wherein each set of the second plurality of sets includes a series of traffic observations each indicating the amount of traffic on the link within an observation interval of a particular duration for that set, a peak value of a traffic observation of the series of traffic observations within each set of the second plurality of sets and an average value of the series of traffic observations within each set of the second plurality of sets, wherein the observation interval durations for each set of the first plurality of sets of traffic observations are equal to the observation interval durations for each set of the second plurality of sets of traffic observations.
4.
A method as claimed in Claim 3, further comprising the steps of:
calculating a second peak to average ratio for each observation interval duration based on the determined peak and average values within each set of the second plurality of sets, wherein the power function is of either the first or second peak to average ratio against its respective observation interval duration.
5. A method as claimed in Claim 4, wherein the power function is of the maximum of either the first or second peak to average ratio against its respective observation interval duration.
6. A method as claimed in any one of Claims 3 to 5, further comprising the step of: determining a peak-utilisation period for the link, wherein the first and second time periods are within the peak-utilisation period.
7. A method as claimed in any one of Claims 3 to 6, further comprising the steps of:
determining first and second overall traffic observations for the first and second time period respectively, wherein the first and second overall traffic observations indicate the amount of traffic on the link in the first and second time periods respectively; and filtering the series of traffic observations of the first and second plurality of sets of any traffic observation that satisfies a criterion based on the first and second overall traffic observations respectively.
8. A method as claimed in any one of the preceding claims, wherein one of the series of observations of a first set of the first and/or second plurality of sets occurs at the same time instance as one of the series of observations of a second set of the first and/or second plurality of sets.
9. A computer program comprising instructions which, when the program is executed by a computer, cause the computer to carry out the method of any one of the preceding claims.
10. A computer-readable data carrier having stored thereon the computer program of Claim 9.
11. A device for dimensioning a link in a telecommunications network, the device comprising a processor adapted to:
determine, for a first plurality of sets of traffic observations, wherein each set of the first
5 plurality of sets includes a series of traffic observations each indicating the amount of traffic on a link within an observation interval of a particular duration for that set, a peak value of a traffic observation of the series of traffic observations within each set of the first plurality of sets and an average value of the series of traffic observations within each set of the first plurality of sets;
10 calculate a first peak to average ratio for each observation interval duration based on the determined peak and average values within each set of the first plurality of sets; determine a goodness of fit value for a power function of the first peak to average ratio against its respective observation interval duration;
select a dimensioning interval duration based on the determined goodness of fit value;
15 and estimate a capacity of the link based on the selected dimensioning interval duration.
Intellectual
Property
Office
Application No: GB1703942.1 Examiner: Mr Jonathan Richards
11. A device for dimensioning a link in a telecommunications network, the device comprising a processor adapted to:
determine, for a first plurality of sets of traffic observations, wherein each set of the first plurality of sets includes a series of traffic observations each indicating the amount of
5 traffic on a link within an observation interval of a particular duration for that set, a peak value of a traffic observation of the series of traffic observations within each set of the first plurality of sets and an average value of the series of traffic observations within each set of the first plurality of sets;
calculate a first peak to average ratio for each observation interval duration based on
10 the determined peak and average values within each set of the first plurality of sets;
determine a goodness of fit value for a power function of the first peak to average ratio against its respective observation interval duration;
select a dimensioning interval duration based on the determined goodness of fit value; and
15 estimate a capacity of the link based on the selected dimensioning interval duration.
09 02 18
Amendments to the claims have been made as follows:
1. A method of dimensioning a link in a telecommunications network, the method comprising the steps of:
5 determining, for a first plurality of sets of traffic observations, wherein each set of the first plurality of sets includes a series of traffic observations each indicating the amount of traffic on a link within an observation interval of a particular duration for that set, a peak value of a traffic observation of the series of traffic observations within each set of the first plurality of sets and an average value of the series of traffic observations within
10 each set of the first plurality of sets;
calculating a first peak to average ratio for each observation interval duration based on the determined peak and average values within each set of the first plurality of sets; determining a goodness of fit value for a power function of the first peak to average ratio against its respective observation interval duration;
15 selecting a dimensioning interval duration based on the determined goodness of fit value;
estimating a capacity of the link based on the selected dimensioning interval duration; and using the estimated capacity to deploy the link in the telecommunications network.
2. A method as claimed in Claim 1, wherein the step of selecting the dimensioning interval duration is based on the goodness of fit value and a service level requirement.
3. A method as claimed in either Claim 1 or Claim 2, wherein the first plurality of
25 sets of traffic observations are within a first time period, and the method further comprises the steps of:
determining, for a second plurality of sets of traffic observations within a second time period, wherein each set of the second plurality of sets includes a series of traffic observations each indicating the amount of traffic on the link within an observation
30 interval of a particular duration for that set, a peak value of a traffic observation of the series of traffic observations within each set of the second plurality of sets and an average value of the series of traffic observations within each set of the second plurality of sets, wherein the observation interval durations for each set of the first plurality of sets of traffic observations are equal to the observation interval durations for
35 each set of the second plurality of sets of traffic observations.
09 02 18
4. A method as claimed in Claim 3, further comprising the steps of: calculating a second peak to average ratio for each observation interval duration based on the determined peak and average values within each set of the second plurality of
5 sets, wherein the power function is of either the first or second peak to average ratio against its respective observation interval duration.
5. A method as claimed in Claim 4, wherein the power function is of the maximum of either the first or second peak to average ratio against its respective observation
10 interval duration.
6. A method as claimed in any one of Claims 3 to 5, further comprising the step of: determining a peak-utilisation period for the link, wherein the first and second time periods are within the peak-utilisation period.
7. A method as claimed in any one of Claims 3 to 6, further comprising the steps of:
determining first and second overall traffic observations for the first and second time period respectively, wherein the first and second overall traffic observations indicate
20 the amount of traffic on the link in the first and second time periods respectively; and filtering the series of traffic observations of the first and second plurality of sets of any traffic observation that satisfies a criterion based on the first and second overall traffic observations respectively.
25 8. A method as claimed in any one of the preceding claims, wherein one of the series of observations of a first set of the first and/or second plurality of sets occurs at the same time instance as one of the series of observations of a second set of the first and/or second plurality of sets.
30 9. A computer program comprising instructions which, when the program is executed by a computer, cause the computer to carry out the method of any one of the preceding claims.
10. A computer-readable data carrier having stored thereon the computer program
35 of Claim 9.
09 02 18
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
GB201703942A GB2560515B (en) | 2017-03-13 | 2017-03-13 | Telecommunications network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
GB201703942A GB2560515B (en) | 2017-03-13 | 2017-03-13 | Telecommunications network |
Publications (3)
Publication Number | Publication Date |
---|---|
GB201703942D0 GB201703942D0 (en) | 2017-04-26 |
GB2560515A true GB2560515A (en) | 2018-09-19 |
GB2560515B GB2560515B (en) | 2019-12-25 |
Family
ID=58605403
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
GB201703942A Active GB2560515B (en) | 2017-03-13 | 2017-03-13 | Telecommunications network |
Country Status (1)
Country | Link |
---|---|
GB (1) | GB2560515B (en) |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2006067769A1 (en) * | 2004-12-23 | 2006-06-29 | Corvil Limited | A method and apparatus for calculating bandwidth requirements |
US20080181125A1 (en) * | 2007-01-31 | 2008-07-31 | Fujitsu Limited | Bandwidth measuring method and device |
WO2017117487A1 (en) * | 2015-12-31 | 2017-07-06 | Echostar Technologies L.L.C | Systems and methods for bandwidth estimation in oscillating networks |
-
2017
- 2017-03-13 GB GB201703942A patent/GB2560515B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2006067769A1 (en) * | 2004-12-23 | 2006-06-29 | Corvil Limited | A method and apparatus for calculating bandwidth requirements |
US20080181125A1 (en) * | 2007-01-31 | 2008-07-31 | Fujitsu Limited | Bandwidth measuring method and device |
WO2017117487A1 (en) * | 2015-12-31 | 2017-07-06 | Echostar Technologies L.L.C | Systems and methods for bandwidth estimation in oscillating networks |
Also Published As
Publication number | Publication date |
---|---|
GB201703942D0 (en) | 2017-04-26 |
GB2560515B (en) | 2019-12-25 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US9763122B2 (en) | System and method for tracking a line rate utilization | |
US9137133B2 (en) | Tracking maximum throughput and service level agreement compliance | |
US8208395B2 (en) | System and method for measuring interface utilization using policers | |
EP4207642A1 (en) | Method and system for guaranteeing service application experience | |
US11171855B2 (en) | Telecommunications network | |
GB2560515A (en) | Telecommunications network | |
US8649484B2 (en) | Apparatus, methods, and articles of manufacture to predict vectored digital subscriber line (DSL) performance gains | |
EP3602967B1 (en) | Network availability measurement aggregation and correlation | |
EP3097660B1 (en) | Methods and apparatus for operating an access network | |
EP3123670B1 (en) | Dynamic line management | |
CN101800666A (en) | Method and system for collecting port information of digital subscriber line | |
Schmidt et al. | Lightweight link dimensioning using sFlow sampling | |
Szymanski et al. | Traffic provisioning in a Future Internet | |
US11043987B2 (en) | Estimating the electrical line length of a digital subscriber line | |
EP3035661B1 (en) | A method for verifying a port synchronisation | |
EP2916532A1 (en) | Dynamic line management | |
CN105101281A (en) | Performance data processing method and device |