EP3417646A1

EP3417646A1 - Cross layer network performance indice for operator benchmarking

Info

Publication number: EP3417646A1
Application number: EP17709367.1A
Authority: EP
Inventors: Siddarth NAIK; Janis NÖTZEL; Eduard Jorswieck
Original assignee: Technische Universitaet Dresden
Current assignee: Technische Universitaet Dresden
Priority date: 2016-02-16
Filing date: 2017-02-16
Publication date: 2018-12-26
Also published as: WO2017140810A1

Abstract

A network device receives information for an infrastructure network, which could include the core network, access network and physical layer and non-physical layer performance indicators. This information could be compared across multiple geographies, access network entities and/or multiple wireless service providers or even multiple regulatory areas. This comparison could help in developing market related indicators, which could be available via a virtual exchange platform to the interested parties in a real-time or non-real time fashion. The network device could then utilize this information for benchmarking the performance of the core and access network entities of a particular operator and across operators. This in turn creates a market value or financial indicators based on wireless indicators.

Description

CROSS LAYER NETWORK PERFORMANCE INDICE FOR OPERATOR BENCHMARKING

Technical Field

The invention relates to a method for benchmarking of operator and network performance of communication

networks .

Background Art

Cellular and non-cellular wireless communications have finite resources over a set of frequency bands, time horizon, energy constraints and monetary constraints.

These resources are shared among multiple users accessing different services simultaneously.

Furthermore, application running different services could create varied demands on the infrastructure of a wireless service provider. This demand needs to be met by the core network and access network infrastructure (which utilize and occupy certain spectrum bands) of the wireless service providers .

A wireless service provider is also known as a mobile network operator, abbreviated as operator which is used in the following. An operator is a provider of services wireless communications that owns or controls all the elements necessary to sell and deliver services to an end user .

Furthermore, physical layer wireless key performance indicators, e.g, signal to interference and noise ratio (SINR) , outage probability, latency in terms of time delay, call/service blocking probability, call/service dropping probability have long been the cornerstones of defining the success of the performance of wireless service. However, there has been very little or no transparency achieved despite the massive technological improvements in the last decades towards bridging the gap between these physical layer wireless engineering

indicators and the financial market or customer accepted performance indicators of a telecommunication company.

Furthermore, the developments towards infrastructure sharing, while having reduced capital expenditure costs for the wireless service providers have added a further layer of opaqueness of how the market value and judges the performance of a particular operator at the following two levels :

1. At each entity level in the core network, noncore

network, radio resource assets and the connection in between them.

2. At an aggregated level, either by aggregation via a. Geography or locations, b. Functional form, e.g. all eNodeBs or all base

stations, or c. Financial aggregation in the form of plant,

property and equipment or licenses and

concessions .

It is the object of the invention to provide indicators for benchmarking mobile network operators. Disclosure of Invention

The object of the invention is solved by a method for benchmarking of operator and network performance,

comprising the steps a. jointly collecting and processing measurements and parameters across multiple cellular communication; b. producing temporal, spectral and spatial

statistics basing on the measurements and

parameters; and c. producing network performance indices (NPI) basing on these statistics; preferably by combining multiple network parameters and measurements at least across different communication layers, time, geographical area and operators to create

indicators for a specific aspect of network performance .

Thereby the method is able to jointly collect and process measurements and parameters across multiple cellular communication layers (similar to layers in open systems interconnections (OSI) model), for producing temporal, spectral and spatial statistics with the purpose of producing network performance indices (NPI). These indices may combine multiple network parameters and measurements across different communication layers, time, geographical area, operators, etc. to create indicators for a specific aspect of network performance.

This information could be compared across multiple

geographies, access network entities and/or multiple wireless service providers or even multiple regulatory areas. This comparison could be available via a virtual exchange platform in a real-time or non-real time fashion.

Example of indices comprise of but are not limited to air interface throughput index, end-to-end throughput index, air interface latency index, end-to-end latency index, air interface utilization index, air interface load index, air interface mobility index, outage index, app coverage index, app outage index, app network performance index (ANPI) .

Examples of app network performance index are Netflix ANPI, UBS mobile ANPI.

In a further embodiment of the invention one or multiple indices are combined to create combined indices.

These ANPI as mentioned above allows the content provider, the network provider and the service provider to jointly optimize the user experience for a particular application and service. Hence, it becomes a much more relevant tracking metric as compared to other more generic metrics.

Furthermore since the RES is the main bottleneck of the network, the metrics could be designed to focus on RBS ' NPIs .

In one particular embodiment of the method the indices are calculated across multiple operators.

Hence, the arrangement could benchmark operators based on these, at a particular geographic granularity and a given pre-specified recurring or non-recurring time horizon.

In one particular embodiment, the performance of various network entities with similar functions within the same operator using the indices. The method could utilize the indices for comparing the performance of various network entities with similar functions (e.g. base stations against base stations) within the same operator. Based on this embodiment the metric could be profitable over a specified time horizon from a third-party investor who is neither the operator, vendor or in any other firm involved in the telecom supply chain except as an

investor . In one particular embodiment, the performance of a

collection of network entities is compared by the indices and the network entities are ranked for a particular operator based on a particular metric.

The indices could be utilized for comparing the

performance of a collection of network entities (either similar or dissimilar) so as to rank them for a particular operator based on a particular metric.

In a further embodiment of the method the various

measurement parameters obtained from various core network, access network components from the various operators are shocked .

A shock is defined as a scaling (usually upscaling of a certain parameter or a set of parameters to visualize its impact on the entire system performance over a given time horizon.

In this embodiment, the method can shock various

measurement parameters obtained from the various core network, access network components from the various operators. Certain examples of what can be shocked are demand function, channels, number of users demanding a particular service.

In a further embodiment all these shocks can be applied, at different localization points and over different time horizons. Examples of some models that can be utilized are e.g. vector error correction model, structural vector auto-regression model, general equilibrium model, dynamic stochastic general equilibrium (DSGE) , panel time series regression .

The method according to the invention offers a lot of advantageous possibilities of usage as following

described . a) These indices could be utilized by the designer (or

requestor) of any network entity to verify if a

particular function of performance indicators is being requested (or designed) as intended by certain pre^¬ arranged agreements. b) The vendors could utilize selection of indices to

define a standard for the network performance given recommended parameters from the vendor.

The operators could then try to optimize their network performance by tuning the network parameters to improve in one or all indices or improve a selected number at the cost of other indices. c) The vendors could utilize these indices to ensure that customers are not utilizing their software in undesired ways. One particular way to misuse the software is systematic reporting of higher interference

temperatures so as to obtain more resources. d) The operators could utilize these indices to benchmark hardware and network devices from various vendors against each other for better future and current resource allocation and expected financial and business strategy planning. e) The regulator could utilize these indices to have a

better handle over the deployment plans of the operator so as to serve its role as an auctioneer of spectrum to the operators who bid the highest price while best serving the end users . f) The vendors and/or operators could utilize these

indices to hedge their operation risk against unwanted movements against that specified in the service level agreements .

Furthermore, for the case of the operators, they could consider reducing their capital expenditures and operating expenditures via benchmarking their network devices as inventory against each other and seeking investors based on the historical performance of these indices .

For example the profits made from customers could be ordered to localized network devices. If the operators offer the network devices as investment vehicles to investors, then a prearranged ratio of profits

collected via this network device could be passed on to the investor. g) The method and arrangement could also have access to the financial layer of the operators, which is

responsible for billing the customer and the enterprise in correspondence to the subscription of the customer or enterprise and their respective consumption. h) The method and arrangement can also have access to

market related financial information with respect to this operator, e.g. credit rating, list of supplier and vendors, list of counter parties for certain asset and liabilities, maturity and volume of the assets and liabilities. i) When operators outsource their network management to subcontractors, indices can be used to concretely define a legal agreement between the parties involved.

These indices could be monitored during the period of the contract and different penalties or bonuses can be defined based on indices. j ) In future scenarios where there maybe companies who run networks like utility companies and rent out their networks to any interested operator or businesses, indices can be used to concretely define and advertise the quality of the network and in the pricing of the offered services. k) The method has the ability to divide the resulting

processed time series (where the input was wireless engineering measurement parameters) into a financial risk and return framework. Hence, the method is able to calculate well established ratio's e.g Sharpe, Sortino and other factors such as alpha and beta when well- defined objects are compared to each other and a benchmark from these well-defined objects is calculated and available.

1) The method may run on a network device, which has the ability to create an order book for user, frequency and resource trading at various geographic granularities.

The objective of the invention is also solved by an arrangement for benchmarking of operator and network performance comprising a network device. The network device is configured to communicate with multiple

operators and to perform multiple calculation taking engineering measurement parameters from the operators and to convert them into performance indicators of the

operators

The network device may be provided with interfaces to financial exchanges and clearing houses. The exchange can then connect to multiple interested counterparties at different granularities and offer the interest parties the appropriate palate of processed engineering measurement parameters for technical benchmarking and financial products .

In an embodiment the arrangement is further comprising a database configured to perform mathematical calculations on physical and non-physical layer measurements of

cellular or non-cellular wireless systems.

In a further embodiment the engineering measurement is arranged in the database configured to perform parameter time series. The time series are mode, median, mean, standard deviation, skew and kurtosis. Furthermore, the database could have the ability to run other well

established mathematical constructs such as K-means, linear regression, non-linear regressions, windsoring etc.

In a further embodiment the database is configured to calculate cross correlation and higher order moments for two or more parameters for influencing an NPI .

In a further embodiment the network device comprises a storage configured to store the time series information of the measurement parameters .

The network device running the method could store the time series information of the measurement parameters for very large time durations via packing it in tar balls for saving it in a compressed format. This data can then be accessed as and when required for constructing historical time series.

In a further embodiment the network device further

comprises an interface between noncore network

infrastructure and the network device, i.e. the network device could have multi-RAT access and off-loading related measurements and information.

In a further embodiment the arrangement is configured to resolve non-synchronous time stamps related data from different wireless operators to build time series data. Standard techniques such as interpolation, e.g. via cubic spline fitting can be utilized for this purpose. Certain loss of data and information might have to be expected based on the particular interpolation method utilized and the quality of the data being received.

In a further embodiment the arrangement is configured to resolve inconsistencies in data quality, e.g.

inconsistencies caused by non-co-located base stations and other geographical inconsistencies, which might need to be resolved, when data from these entities are stored in a database . Modes for Carrying Out the Invention

In the following the invention is described by examples.

Fig. 1 shows a schematic overview of an arrangement

according to the invention performing the

inventive method and

Fig. 2 shows an abstraction of an example applying the technical invention for commercial purposes.

As shown in Fig. 1 arrangement 1 for benchmarking of operator and network performance comprises a network device 2. The network device 2 is configured to

communicate with multiple operators 3 and to perform multiple calculation taking engineering measurement parameters from the operators 3 and to convert them into performance indicators 4 of the operators 3.

The network device is provided with interfaces 4 to the financial layer 8, e.g. financial exchanges and clearing houses .

The operators comprise a core equipment 5 and a non-core equipment 6.

As shown in Fig. 2 end users 8 consume voice and data services from the telecommunication operator 3 shown with arrow 9. For their consumption the make payments for voice and data services, shown with arrow 10.

On the other side investors 11 subsidize capital

expenditures of operators by buying infrastructure bonds in bond auctions, as shown with arrow 12. In return the operator 3 passes through percentage of payments as floating coupon payments to investors 11 for baerig the funding risk of the infrastructure. The calculation of the coupon payments is done via the processing of information received by the network device 2, i.e. by the performance indicators 4.

The commercial usage of the invention could be further described with the following examples

Berlin-Munich Outage exchange traded fund (ETF) : The network device 2 considers the interstate highway between the two German cities of Berlin and Munich and all the geographic locations occurring between them as the source of data. It then builds a daily time series of all the outages occurring on this interstate highway and stores them classified by service class, air interface (e.g. GSM, EDGE, GPRS, UMTS, LTE) and different time-buckets.

A time bucket is a particular time horizon as seen from the point of observation and or calculation, namely at the present point in time. A time buckets could be for example one day, one week, two weeks, one month, three months, six months, 12 months.

The price of the ETF is some function of a particular classification of outages. As an example, if there are IOOK voice calls made via UMTS and LTE each day on this interstate highway and IK of these faced outage, then the ETF could be priced at for example as function of (a x 100,000 - b x 1, 000) /l, 000. We can choose a = 1 and b = 1, resulting in 99. Hence, as the number of outages

increases, the value of outage ETF decreases. The ETF can be marked to the wireless engineering indicators

(analogous to mark to model) on a periodical basis. In the example just shown, this can be done once a day. Shorter time horizon, e.g. one minute, one second could be initiated based on the demand for the required risk class.

One can easily observe similarities between an outage ETF and a Short term interest rate (STIR) Futures product. For the STIR Futures product as the interest rate increases, the price of the Futures product decreased. Hence, outage plays the role of interest rates. However, the concept of negative outage does not seem to be realistic (maybe we need to extend the definition of outage) .

Frankfurt outage probability insurance note: In one particular embodiment of the method and apparatus, the device considers the city of Frankfurt in Germany as the geography for measuring outage probabilities at all localization points as a time series. In this particular case the network device transmits the outage probability of all entities in Frankfurt from a particular operator, e.g. Vodafone, to an insurance firm, e.g. Munich Re. The operator usually would require a certain monetary budget in the case there is a very large outage caused by factors outside of its control (which would have to legally documented and accepted by the insurance firm) to account for uninterrupted network service. The operator can buy an insurance note specifically for this purpose. Along with other financial and legal drivers, one of the main drivers being fed into the triggering or non-triggering event for the insurance note can be the outage probabilities at regular time intervals processed and provided the network device. As in most classical insurances, the operator would pay a periodic premium to the insurance firm and the insurance firm would disburse the agreed upon insurance sum based on the inputs from the network device regarding outage events along with other legal and financial clauses agreed between these two concerned parties. Collateralized outage obligation for an eNodeB or for a set of eNodeBs in a geographic location: For the purpose of funding the premiums from the previous point, the operator could create collateralized outage obligations (COO) . These COO would be physical infrastructure

performance based financial instruments, whose market values will be based on the data provided to the network device .

Paradeplatz and vicinity, Zurich based mobile payment based cash flow bonds: Various base stations could be classified on their performance and the number of cash flows they generate for the operators on a daily basis. Mapping of NPI units into unit of monetary reimbursement or units of service reimbursement (e.g. units of

guarantees of throughput) at a future point in time would need to be drilled down to their contribution at an RBS level via information about total usage, subscription and other parameters, which can be accessed from the core network .

These units of monetary reimbursement or units of service reimbursement can be segregated by the data and mobility influencing these base stations.

The data and mobility models affecting these base stations directly influences their profitability. Based on the performance indicators of the base stations and their profitability from revenues generated from voice and data on a minute basis, hourly basis, daily basis, weekly basis, monthly basis etc., they can be ranked into

different grades. As against telecommunication companies paying for all the infrastructure costs as front loaded capital expenditure for this non-core equipment and the related periodic utility bills and maintenance, the telecommunication companies could auction bonds on these infrastructure equipment. This would significantly reduce their capital expenditure costs. Furthermore percentage values of cash flows generated from the end users are passed on to the bond holders.

The network device helps regulate and verify that the stochastic cash flows or coupons are in accordance with what was promised to the bond holders. The bond holders are able to get exposure to a tailor made risk set of idiosyncratic risk of these localized base stations as against the entire telecommunication operator balance sheet .

The network device and the data processed by it plays a key role in the structuring of these contracts.

Munich Outage probability based Lotto: Retail investors can individuals can bet on the aggregated outage

probability in Munich being above or below a certain threshold. This will create further transparence between the operators performance and help compare operators to each other. Furthermore, the individuals have a mechanism to benefit from the irregularities from the

telecommunication operators performance.

The network device can be a source of such data and verify that all legal and regulatory conditions are met by the operator, when offering such a Lotto. eNodeB non-functioning linked security and SINR

performance based guarantees are other examples of where the network device described in this apparatus could assist in providing accurate information. Method and Arrangement for Benchmarking of Operator and Network Performance of communication networks

List of Nomenclature and Acronyms

Node A cellular network entity defined in a

standardization for a given cellular network technology. Examples of nodes in 3GPP standardization for third generation mobile technology are NodeB, RNC, GGSN .

RBS Radio Base Station

RNC Radio Network Controller

SINR Signal to Interference and Noise Ratio KPI Key Performance Indicators

QoS Quality of Service

GSM Global System for Mobile Communications

EDGE Enhanced Data rates for GSM Evolution

GPRS General Packet Radio Service UMTS Universal Mobile Telecommunications Systems

LTE Long Term Evolution

UE User equipment

UTRA UMTS Terrestrial Radio Access E-UTRA Enhanced UTRA eNodeB E-UTRAN Node B, also known as Evolved Node B,

(abbreviated as eNodeB or eNB) is the element in E-UTRA of LTE that is the evolution of the element Node B in UTRA of UMTS, It is the hardware that is connected to the mobile phone network that

communicates directly with UEs, like a base station in GSM networks.

List of Reference Numbers 1 arrangement

2 network device

3 telecommunication operator

4 performance indicator

5 interface

6 core equipment (core network)

7 no-core equipment (non-core network)

8 user

9 arrow for user' s consumption

10 arrow for user's payment

11 investor

12 arrow for investor' s payment

13 arrow for investor's return

Claims

Method for Benchmarking of Operator and Network

Performance, comprising the steps a. jointly collecting and processing measurements and parameters across multiple cellular communication; b. producing temporal, spectral and spatial

statistics basing on the measurements and

indicators for a specific aspect of network performance .

Method according to claim 1, wherein the indices are comprising at least one of the following indices: air interface throughput index, end-to-end throughput index, air interface latency index, end-to-end latency index, air interface utilization index, air interface load index, air interface mobility index, outage index, app coverage index, app outage index, app network performance index (ANPI) .

3. Method according to claim 2, wherein the app network performance index (ANPI) is comprising Netflix ANPI, and/or UBS mobile ANPI.

4. Method according to claim 2 or 3 wherein one or

multiple indices are combined to create combined indices .

5. Method according to at least one of the claims 1 to 4 wherein the indices are calculated across multiple operators .

6. Method according to at least one of the claims 1 to 5 wherein the performance of various network entities with similar functions within the same operator using the indices.

7. Method according to at least one of the claims 1 to 6 wherein the performance of a collection of network entities is compared and the network entities are ranked for a particular operator based on a particular metric .

8. Method according to at least one of the claims 1 to 7 wherein the various measurement parameters obtained from various core network, access network components from the various operators are shocked.

9. Method according to claim 8 wherein the shocks are

applied at different localization points and over different time horizons.

10. Arrangement for benchmarking of operator and network performance, comprising a network device wherein the network device is configured to communicate with multiple operators and to perform multiple calculation taking engineering measurement parameters from the operators and to convert them into performance

indicators of the operators

11. Arrangement according to claim 10 further comprising a database configured to perform mathematical

calculations on physical and non-physical layer

measurements of cellular or non-cellular wireless systems .

12. Arrangement according to claim 11 wherein the

engineering measurement is arranged in the database configured to perform parameter time series.

13. Arrangement according to at least one of the claims 10 to 12 wherein the database is configured to calculate cross correlation and higher order moments for two or more parameters for influencing an NPI .

14. Arrangement according to at least one of the claims 10 to 13 wherein the network device comprises a storage configured to store the time series information of the measurement parameters .

15. Arrangement according to at least one of the claims 10 to 14 wherein the network device further comprises an interface between noncore network infrastructure and the network device.

16. Arrangement according to at least one of the claims 10 to 15 configured to resolve non-synchronous time stamps related data from different wireless operators to build time series data.

17. Arrangement according to at least one of the claims 10 to 16 configured to resolve inconsistencies in data quality, e.g. inconsistencies caused by non-co-located base stations and other geographical inconsistencies.

Arrangement according to at least one of the claims 10 to 17 further comprising an interfaces to multiple network devices so as to aggregate measurement

parameters across multiple layers, across multiple geographies, financial entities, functional entities or wireless operators.