METHOD FOR DETERMINING A NETWORK CARBON FOOTPRINT OF AT LEAST ONE DEVICE AND AN ASSOCIATED STRUCTURE

Abstract

1. Method for determining a carbon footprint (Ec) of a device comprising the following steps: - retrieving (30) a set of network flows (13); - determining (31), for each network flow, a set of sections (R1-R5); - searching (32), in a database (19), whether one or more detected sections are already associated with a carbon cost; if no carbon cost is referenced for a detected section: - geolocation (34) of interconnection nodes of the section (R1-R5); - projection (35) of the section between said two interconnection nodes; - estimation (36) of a distance of the section in one or more countries; and - incrementing (37) said carbon footprint (Ec) with a carbon cost per distance per country multiplied by the distance of the section in each country; the carbon footprint (Ec) being obtained following all the incrementing steps for the different sections. Figure for short: Fig 5.

Description

Title of the invention: METHOD FOR DETERMINING A NETWORK CARBON FOOTPRINT OF AT LEAST ONE PER IPHERICAL AND ASSOCIATED STRUCTURE Field of invention

[0001] The invention relates to the technical field of measuring the digital carbon footprint, i.e. measuring the carbon footprint associated with digital data exchanges carried out on the Internet network.

[0002] More particularly, the invention aims to estimate the carbon footprint of a device based on a set of network flows associated with this device.

[0003] By studying several peripherals of a structure, the invention also aims to estimate the network carbon footprint of an entire structure comprising several peripherals.

[0004] The invention makes it possible to characterize the carbon footprint of the network flows of a device or a structure in order to allow an individual, a company or a community to adopt strategies for limiting the carbon footprint. Thus, the invention can be used in a large number of fields for which network exchanges are implemented by digital devices, such as computers, smartphones or even connected objects. State of the art

[0005] Climate change now requires a massive reduction in anthropogenic pollution. To achieve this, all individuals, businesses, and communities are being asked to contribute to finding ways to reduce the pollution generated by their activities.

[0006] For industries, pollution reduction strategies are often linked to the manufacturing process or product logistics. For individuals, service companies, or communities, the use of digital tools generally plays a key role in the carbon footprint of their business.

[0007] However, today, it remains very difficult to characterize the carbon footprint of a structure that mainly uses digital tools. In particular, digital activities generate a large number of network exchanges involving many components, mainly routers, network connection nodes and servers, for which it is difficult to estimate the carbon cost associated with the use of this equipment.

[0008] It is therefore practically impossible for a structure to determine today what the carbon footprint associated with its digital activity is; so much so that it is difficult to define a strategy to reduce its digital carbon footprint.

[0009] For example, if a company needs to renew servers, it can choose to buy servers to place them within the company or to use decentralized servers accessible via the Internet. In the first case, it is possible to research the carbon footprint of manufacturing and powering the servers for their entire lifespan. In the second case, it is also possible to research the carbon footprint of manufacturing and powering the servers for their entire lifespan, but it is also necessary to consider the increase in network flows to reach the decentralized servers. However, there is currently no known solution for estimating the carbon footprint of these network flows.

[0010] The French Agency for Ecological Transition (ADEME) attempted to address this challenge in a July 2011 report entitled "Comparative Analysis of the Environmental Impacts of Electronic Communication." In this report, ADEME estimates the carbon footprint of an email exchange by calculating the carbon footprint of using a sending computer, a first transmission network, a data center, a second transmission network, and a receiving computer. Given the many elements to be taken into account, a large number of approximations are made, particularly on the transmission networks. It follows that the result of this report is independent of the distance between the sending computer, the data center, and the receiving computer.

[0011] Thus, this study does not allow us to estimate the carbon footprint of a structure's network flows, for example to estimate the difference between internal or external data storage or between two external storages hosted in different locations.

[0012] There is therefore always a technical problem in estimating the carbon footprint of network exchanges of a device or a set of devices in a structure in relation to IT asset management strategies. Furthermore, another IT asset management strategy may correspond to an authorization or not to use certain functionalities for users.

[0013] It follows that many strategic decisions are taken today within structures without considering the digital carbon footprint. Statement of the invention

[0014] To address this technical problem, the invention proposes to trace several network flows of a device in order to obtain the position of the nodes of each network flow, each network flow passing through several sections between the different nodes. It is thus possible to calculate the distance traveled for each section, in each country.

[0015] By making the approximation that the carbon footprint of these network flows depends on the energy mix and the distance traveled in each country, it is possible to associate a carbon cost with each section. By aggregating the different carbon costs of the different sections of several network flows of a device, it is possible to estimate the carbon footprint of a device.

[0016] This carbon footprint is thus correlated to the IT management strategy of a structure so that if a company plans to use an internal server or a decentralized server, it is possible to test the impact on a device in order to estimate the carbon footprint that this strategy would have on the entire structure.

[0017] Thus, according to a first aspect, the invention relates to a method for determining a carbon footprint of a device comprising the following steps: - recovery of a set of network flows associated with said device; - determination, for each network flow, of a set of sections making it possible to connect said peripheral with a transmitter or a receiver of the network flow, each section being connected by interconnection nodes; - searches, in a database, if one or more detected sections are already associated with a carbon cost; if a carbon cost of a detected section is recorded in said database: - incrementing said carbon footprint with the carbon cost associated with the section recorded in said database; if no carbon cost is referenced for a detected section: - geolocation of the interconnection nodes of the section; - projection of the section between said two interconnection nodes; - estimation of a distance of the section in one or more countries; and - increment of said carbon footprint with a carbon cost per distance per country multiplied by the distance of the section in each country; the carbon footprint of a device being obtained following all the incrementation steps for the different detected sections of the different network flows considered on the set of network flows.

[0018] The invention therefore proposes to make an approximation according to which the carbon footprint of a section is linked to the distance traveled in each country, so that the energy mix of the country makes it possible to account for this carbon footprint.

[0019] To do this, a first step consists of retrieving a set of network flows associated with a device. These network flows can be retrieved by also capturing a protocol and / or a port used for each network flow so that the carbon footprint of a device can be obtained for each protocol and / or each port captured in the set of network flows associated with said device. Thus, it is possible to measure the impact on the carbon footprint of activating or deactivating certain functionalities for users. For example, certain Video streaming services use TCP port 1935. By looking at the carbon footprint of all devices in a structure on this TCP port 1935, it is possible to estimate the gain obtained by restricting the functionality of watching movies in streaming for users.

[0020] The network flows studied for a device may be all the network flows or a sample of network flows taken at random or according to a predetermined strategy. For example, network flows presenting a large volume of data may be considered in a predominant manner over others.

[0021] In this embodiment, the step of recovering a set of network flows associated with said device is carried out on all the data transmitted or received by the device with a data volume greater than a threshold value.

[0022] This embodiment makes it possible to limit the analysis time of network flows by mainly considering those which generate processing time and therefore significant consumption for routers, network connection nodes and servers.

[0023] Furthermore, these network flows may correspond to real network flows captured on a network or to simulated network flows.

[0024] To capture real network flows, it is for example possible to use an agent, a monitoring device or a module for recovering a firewall log.

[0025] According to a first variant, the step of recovering a set of network flows associated with said device is carried out by means of an agent installed on the device and configured to capture the network exchanges of the device. The agent may correspond to a Microsoft Windows service or a Linux daemon installed on an operating system of a device. With this embodiment, it is sufficient to deploy software on a device to recover the set of network flows.

[0026] According to a second variant, the step of recovering a set of network flows associated with said peripheral is carried out by means of a supervision device connected to a network of the peripheral. The supervision device may correspond to a box integrating one or more network ports and a circuit for capturing network exchanges on an internal network of a structure. This embodiment makes it possible to implement the invention without modifying the configuration of the peripherals of a structure.

[0027] According to a third variant, the step of recovering a set of network flows associated with said device is carried out by means of a module for recovering a log of a firewall configured to trace all the network flows of a structure. This embodiment proposes to implement the invention by simply analyzing the log file of the network exchanges of a structure.

[0028] In addition to these three variants, it is also possible to combine agents with a supervision device to carry out, with the recovery of network flows, several measurements simultaneously on several devices.

[0029] Whatever the embodiment, with an agent, a monitoring device or a module for recovering a log from a firewall, it is possible to carry out the invention directly at the level of the device, the monitoring device or the firewall. As a variant, all the network flows to be analyzed can be transmitted to a server configured to estimate the carbon footprint of the device.

[0030] With regard to the nature of the network flows, an agent, a monitoring device or a module for retrieving a log of a firewall can capture real network flows and the invention can be implemented directly on the basis of these real network flows. Alternatively, it is possible to simulate expected network flows on the basis of these real network flows. For example, by detecting that a device carries out approximately 10,000 network flows per day to a file server, it is possible to estimate the impact of a move of the file server by carrying out a simulated network flow.

[0031] To do this, the simulated network flows can be generated using one or more “ping” commands on the server under consideration to determine the impact on the carbon footprint of changing servers with the different routes that the different network flows can take.

[0032] In a second step, each network flow considered, real or simulated, is studied to obtain the set of sections making it possible to connect the device with a transmitter or a receiver of the network flow. To do this, it is possible to use the network diagnostic tool "traceroute" or "tracert". This network diagnostic tool makes it possible to follow the path taken by a network flow between a transmitter and a receiver by describing all the routers and nodes used. With this diagnostic tool, each node is conventionally described by its IP address, which can be geo-located. This second step of determining the nodes and sections can be carried out by the agent, the monitoring device or the module for retrieving a log from a firewall in order to be transmitted with the network flows to a server configured to carry out the estimation of the carbon footprint of the device, in particular the geolocation of the nodes.

[0033] Before carrying out this geolocation, the invention proposes searching in a database whether the carbon cost of a section is already known, thus making it possible to limit the overall analysis time when sections are already known from the database.

[0034] When the carbon cost of a section is not known from the database, the invention proposes to geolocate the nodes of the section to estimate the distance of the section. To do this, a projection is carried out between the nodes. For example, this projection can correspond to a projection of a line following the curvature of the earth.

[0035] Following the projection of the section between two nodes, it is possible to determine whether the section extends over one or more countries. For example, it is possible to consider that a section extends over a single country if the entire section extends over the same country. To do this, one embodiment proposes sampling the section at a set of points, for example one point every 10 km of the projection, each point of the set of points being geolocated to determine at least the country in which the point is located. Thus, it can be considered that the section extends in a single country if all the points of the section are positioned in the same country, even if the section crosses a border between two points twice.

[0036] If several points are positioned in different countries, it is necessary to determine the position of the border to estimate the distance traveled by the section and by country. To do this, one embodiment proposes to make the approximation according to which a border is always located equidistant between two points.

[0037] In this embodiment, when two points of the set of points are located in distinct countries, the position of a border is estimated equidistant between the points. For example, if the distance between two points is set to 20 km, when two points will be detected in different countries, it is possible to consider that 10 km are traveled in each country to reach these points without precisely searching for the position of the border and the actual distance traveled to the border to limit the calculation time.

[0038] With the distance traveled by a section per country, it is then possible to determine the carbon footprint of the section by multiplying a carbon cost per distance per country with the distance of the section in each country.

[0039] In one embodiment, before performing the step of incrementing said carbon footprint with a carbon cost per distance per country multiplied by the distance of the section in each country, the method implements a step aimed at detecting whether a specific network link is used for the section and, if so, the step of incrementing said carbon footprint is performed according to the expected cost of the specific network link detected. For example, if the position of two nodes is located on each side of the Atlantic Ocean, the section typically comprises a transatlantic optical cable whose consumption is generally known.

[0040] If the carbon cost of this section is not known from the database, it is possible to consider that the carbon cost is similar to the other transatlantic cables whose carbon cost is known.

[0041] This embodiment therefore makes it possible to simplify the calculations and improve their precision because the carbon cost per country is not always relevant when links specific networks are used, such as tunnels or transatlantic or transpacific cables.

[0042] When the carbon cost of a section has been calculated, this carbon cost is preferentially recorded in the database to limit the future analysis time for this section. In this embodiment, the step of incrementing said carbon footprint with a carbon cost per distance per country multiplied by the distance of the section in each country also records the carbon cost in the database.

[0043] Furthermore, the database is preferably updated periodically, for example every month. To do this, the carbon cost of a section can be recorded in the database with an expiry period so that, when this expiry period is reached, the carbon cost is deleted. Thus, the carbon costs are updated regularly to adapt to variations in the allocation of public IP addresses of countries or to changes in the energy mix.

[0044] Finally, the carbon footprint of a device is determined by accumulating the different carbon costs of the different sections of the different network flows considered for this device.

[0045] In addition, it is also possible to search for the network footprint of a structure by accumulating several carbon footprints of several devices of a structure.

[0046] Thus, according to a second aspect, the invention relates to a method for determining a carbon footprint of a structure comprising a set of peripherals, the carbon footprint of the structure being determined by accumulating the carbon footprint of the set of peripherals according to the first aspect of the invention. Summary description of the figures

[0047] The manner of carrying out the invention as well as the advantages which result therefrom will emerge clearly from the following embodiments, given for informational but non-limiting purposes, with the support of the figures in which:

[0048] [Fig.l] [Fig.l] is a schematic representation of an information system associated with a unit for determining a carbon footprint of a peripheral according to a first embodiment of the invention;

[0049] [Fig.2] [Fig.2] is a schematic representation of an information system associated with a unit for determining a carbon footprint of a device according to a second embodiment of the invention;

[0050] [Fig.3] [Fig.3] is a schematic representation of an information system associated with a unit for determining a carbon footprint of a device according to a third embodiment of the invention;

[0051] [Fig.4] [Fig.4] is a schematic representation of a network flow using several segments crossing several countries; and

[0052] [Fig.5] [Fig.5] is a flowchart of the steps for determining a carbon footprint of a device according to one embodiment of the invention. Detailed description of the invention

[0053] [Fig.l] illustrates an information system 10, for example an information system of a structure such as a company or a community. The information system 10 comprises a firewall 11, a router 12, computers and other peripherals 14, such as smartphones, copiers or servers. Thus, the information system 10 is composed of a set of diverse and varied peripherals 14 operating with diverse hardware and software means. These peripherals 14 are connected to a network or a plurality of subnetworks connected to each other.

[0054] To implement the invention, it is necessary to recover a set of network flows associated with a device or a set of devices 14. To do this, a device of interest can integrate an agent 15, as illustrated in [Fig.l].

[0055] The agent 15 preferably corresponds to software means, for example a Microsoft Windows service or a Linux daemon. The agent 15 is configured to capture the network exchanges of the device 14 on which it is installed. Furthermore, the agent 15 can also determine, for each network flow, a set of sections R1-R5, for example by means of the network diagnostic tool “traceroute” or “tracert” which makes it possible to know all the interconnection nodes N1-N6 and the sections R1-R5 of a network flow.

[0056] As illustrated in [Fig.4], a network flow may correspond to an exchange between Madrid and Reykjavik. In this network flow, a first segment RI connects Madrid to Toulouse, a second segment R2 connects Toulouse to Paris, a third segment R3 connects Paris to London, a fourth segment R4 connects London to Manchester and a fifth segment R5 connects Manchester to Reykjavik. In this example, the nodes NI, N2, N3, N4, N5 and N6 are respectively located in Madrid, Toulouse, Paris, London, Manchester and Reykjavik.

[0057] When all the nodes N1-N6 and the sections R1-R5 are determined, the agent 15 transmits a file 13 integrating the details of all the network flows of its peripheral 14 to a server 17 configured to carry out the estimation of the carbon footprint Ec of the peripheral 14. This server 17 is accessible via the Internet 18 and it communicates with a database 19 integrating carbon costs associated with known sections.

[0058] Alternatively, the file 13 integrating the details of all the network flows of a device 14 can be obtained by a supervision device 16, as illustrated in [Fig.2]. The supervision device 16 can correspond to a physical box integrating a motherboard, a processor, memory and a network card.

[0059] In another variant, the file 13 integrating the details of all the network flows of a device 14 can be obtained by a module for retrieving a log installed on the firewall 11, as illustrated in [Fig.3].

[0060] The agent 15, the supervision device 16 or the recovery module installed on the firewall 11 can be configured to recover all or part of the network flows from a device 14 or a set of devices 14. For example, only the network flows with a data volume greater than a threshold value are captured.

[0061] As illustrated in [Fig.5], the server 17 is for example configured to implement steps 30 to 38 in order to estimate the carbon footprint Ec of a device 14. The first step 30 consists of recovering the network flows of the device 14, typically by means of the file 13. The second step 31 aims to determine, for each network flow, the different interconnection nodes N1-N6 and the different sections R1-R5.

[0062] This information can also be extracted from file 13 by considering, in the algorithm of [Fig.5], the sections one by one, that is to say by returning to step 31 when the carbon cost of a section is obtained R1-R5 and incremented with the carbon cost previously calculated.

[0063] Obtaining the carbon cost of a section implements a step 32 of searching the database if the section considered is already associated with a known carbon cost. In the example of [Fig.5], step 32 does not allow the carbon cost Cc(Rl) of the first section RI to be extracted because this carbon cost Cc(Rl) is not known from the database 19. A geolocation step 34 of the interconnection nodes NI and N2 of the first section RI is then carried out. In this example, the IP address of the node NI is 90.175.16.145 and it corresponds to the GPS position: 40.4167; -3.704 while the IP address of the node N2 is 89.86.9.235 and it corresponds to the GPS position: 43.604; 1.443.

[0064] The nodes NI and N2 being located in different countries, Spain and France, it is therefore appropriate to search for the distance of the section RI in each country. To do this, a projection of the section RI is calculated, in a step 35, between the two nodes NI and N2 whose GPS position is known from the geolocation step 34. Preferably, this projection is carried out by following the curvature of the Earth with a discretization every 10 or 20 km, so as to obtain a set of projection points. It is thus possible to determine the position of each of the points of the projection and to estimate the position of the border Fl between the points located in different countries. To simplify the calculations, it is preferable to consider that the border Fl is located equidistant between the points of the different countries. For example, the border Fl can be estimated at the GPS position 42.542; 1.967.

[0065] From the position of the border F1 and the two nodes N1 and N2, it is possible to estimate the distance traveled by the section RI in each country, at step 36. In the example of [Fig.5], the distance traveled by the section RI on Spain D(N1-F1) is 500 km while the distance traveled by the section RI on France D(F1-N2) is 123 km.

[0066] From the distance traveled per country for a section, step 37 describes the incrementation of the carbon footprint Ec with the carbon cost per distance per country multiplied by the distance traveled per country.

[0067] Typically, considering that the energy mix of Spain induces a carbon cost per distance C(ES) / km of 0.15 mgCo2 / km and that the energy mix of France induces a carbon cost per distance C(FR) / km of 0.08 mgCo2 / km, the RI section has a carbon cost of 500*0.15+123*0.08, or 84.84 mgCo2. It should be noted that this figure of 84.84 mgCo2 for the RI section is only given as an explanation of the invention but does not correspond to reality. Furthermore, the carbon cost per distance and per country can be updated periodically depending on the country's actual production over time and the evolution of the energy mix.

[0068] When the carbon cost of the section RI is determined, the carbon footprint Ec is incremented by this carbon cost in step 37. If the section RI is the first section calculated, the carbon footprint Ec therefore becomes equal to 84.84 mgCo2. Furthermore, step 37 can also proceed to record this carbon cost of the section RI in the database 19 to limit the future analysis time for this section. Preferably, the carbon cost of a section is recorded in the database 19 with an expiry period so that, when this expiry period is reached, the carbon cost is deleted. Furthermore, the database 19 can also be completely deleted periodically or following major events, for example a disaster which globally modifies the energy mix of many countries.

[0069] The method then continues with the analysis of the following sections R2-R5 by returning to step 31. In the example of [Fig.5], step 32 makes it possible to directly extract the carbon cost Cc(R2) and Cc(R3) of the sections R2 and R3. For example, the carbon cost Cc(R2) can be equal to 12.3 mgCo2 and the carbon cost Cc(R3) can be equal to 8.2 mgCo2. When the carbon cost of a section is known, it is directly incremented to the carbon footprint Ec in step 33.

[0070] Thus, at the end of the two steps 33 for sections R2 and R3, the carbon footprint Ec therefore becomes equal to 84.84 mgCo2 + 12.3 mgCo2 + 8.2 mgCo2, or 105.34 mgCo2.

[0071] As for section R4, this is not known to database 19 but is located in a single country: the United Kingdom. Thus, after steps 34 and 35 of geolocation and projection, it is sufficient to estimate the distance between the two geolocations of nodes N4 and N5, in step 36, to obtain a single distance traveled. In the example of [Fig.5], this distance is 262 km. Considering that the energy mix of the United Kingdom induces a carbon cost per distance of 0.2 mgCo2 / km, the R4 section has an estimated carbon cost of 0.2*262, or a carbon cost of 52.4 mgCo2. At step 37 of calculating the R4 section, the carbon footprint Ec therefore becomes equal to 105.34+52.4, or 157.74 mgCo2.

[0072] For some sections, it may be interesting to search whether these sections use a specific network link. Typically, section R5 conventionally uses a tunnel T1 connecting the tip of the United Kingdom to Iceland. In a similar manner to the search for the border F1 of section RI, the position of the entrance F2 of the tunnel T1 can be searched in step 35. For example, the entrance F2 of the tunnel T1 can be geolocated at the GPS position 50.906; -3.474. Considering that the tunnel T1 has a fixed carbon cost of 0.01 mgCo2, it is therefore appropriate to search for the carbon cost between the node N5 and the entrance F2 of the tunnel T1 as well as the carbon cost between the exit of the tunnel T1 and the node N6.

[0073] In the example of [Fig.5], the distance between node N5 and entrance F2 of tunnel T1 is estimated at 450 km while the distance between the exit of tunnel F1 and node N6 is estimated at 13 km. Considering that the energy mix of the United Kingdom induces a carbon cost per distance of 0.2 mgCo2 / km and that the energy mix of Iceland induces a carbon cost per distance of 0.05 mgCo2 / km, section R5 has a carbon cost of 450*0.2+13*0.05+0.01, or 90.66 mgCo2. At step 37 of calculating section R5, the carbon footprint Ec therefore becomes equal to 157.74 +90.66, or 248.4 mgCo2.

[0074] Step 38 thus makes it possible to extract the carbon footprint Ec for a device 14 and a single network flow at 248.4 mgCo2.

[0075] Considering a company based in Madrid and which uses a file server located in Reykjavik, we can therefore expect that the digital carbon footprint of the company will be greatly impacted by the multiplicity of network exchanges between these two points. If the company plans to move its servers to Paris, we can expect that the network exchanges between the company and the servers will follow only the RI and R2 sections. Thus, for each network flow, we can estimate that the carbon cost of accessing the file server would go from 248.4 mgCo2 to 97.14 mgCo2 (84.84 mgCo2 + 12.3 mgCo2). A company that would use this type of information could therefore optimize its network architecture to minimize its overall carbon footprint.

[0076] Although the example in [Fig.5] is given for a device 14 and for a single network flow, it is possible to aggregate several flows coming from one or more devices. spherical 14 without changing the invention.

[0077] It follows that it is possible to obtain the carbon footprint of a structure 10 comprising a set of peripherals 14 by considering the carbon footprint Ec calculated on all or part of the peripherals 14.

[0078] Furthermore, the network flows 13 can also be recovered by also capturing a protocol and / or a port used for each network flow. Thus, the carbon footprint Ec of a device 14 or of a structure 10 can be obtained for each protocol and / or each port captured in the set of network flows 13. It is thus possible to anticipate the impact of the limitation of the functionalities for the users, for example if certain protocols or certain ports are blocked on the computers of the employees of a service company.

[0079] The invention thus makes it possible to provide metrics for limiting the digital carbon footprint, i.e. the carbon footprint of the network exchanges of a device or a set of devices of a structure in relation to IT park management strategies.

Claims

Claims

1. Method for determining a carbon footprint (Ec) of a device (14) comprising the following steps: - recovery (30) of a set of network flows (13) associated with said peripheral (14); - determination (31), for each network flow, of a set of sections (R1-R5) making it possible to connect said peripheral (14) with a transmitter or a receiver (20) of the network flow, each section (R1-R5) being connected by interconnection nodes (N1-N6); - search (32), in a database (19), if one or more sections (R1-R5) detected is already associated with a carbon cost; if a carbon cost of a detected section (R1-R5) is recorded in said database (19): - increment (33) of said carbon footprint (Ec) with the carbon cost (Cc(R2), Cc(R3)) associated with the section (R1-R5) recorded in said database (19); if no carbon cost is referenced for a detected section (R1-R5): - geolocation (34) of the interconnection nodes (N1-N6) of the section (R1-R5); - projection (35) of the section (R1-R5) between said two interconnection nodes (N1-N6); - estimation (36) of a distance of the section (R1-R5) in one or more countries; and - incrementation (37) of said carbon footprint (Ec) with a carbon cost per distance per country (C(ES) / km, C(FR)km) multiplied by the distance of the section in each country (D(Nl-Fl), D(F1-N2)); the carbon footprint (Ec) of a device (14) being obtained following all the incrementation steps (33, 37) for the different sections (R1-R5) detected from the different network flows considered on the set of network flows (13).

2. Method for determining a carbon footprint of a device according to claim 1, in which the step of retrieving (30) a set of network flows (13) associated with said device (14) is carried out by also capturing a protocol and / or a port used for each network flow so that the carbon footprint (Ec) of a device (14) can be obtained for each protocol and / or each port captured in the set of network flows (13) associated with said device spherical (14).

3. Method for determining a carbon footprint of a device according to claim 1 or 2, in which the step of recovering (30) a set of network flows (13) associated with said device (14) is carried out on all the data transmitted or received by the device with a data volume greater than a threshold value.

4. Method for determining a carbon footprint of a device according to one of claims 1 to 3, in which the step of recovering (30) a set of network flows (13) associated with said device (14) is carried out by means of an agent (15) installed on the device (14) and configured to capture the network exchanges of the device (14).

5. Method for determining a carbon footprint of a device according to one of claims 1 to 3, in which the step of recovering (30) a set of network flows (13) associated with said device (14) is carried out by means of a supervision device (16) connected to a network of the device (14).

6. Method for determining a carbon footprint of a device according to one of claims 1 to 3, in which the step of recovering (30) a set of network flows (13) associated with said device (14) is carried out by means of a module for recovering a log of a firewall (11) configured to trace all the network flows of a structure.

7. Method for determining a carbon footprint of a device according to one of claims 1 to 6, in which the step of projection (35) of the section (R1-R5) between said two interconnection nodes (N1-N6) carries out a sampling of the section according to a set of points, each point of the set of points being geolocated to determine at least the country in which the point is located.

8. A method for determining a carbon footprint of a device according to claim 7, wherein when two points of the set of points are located in distinct countries, the position of a border (F1, F2) is estimated to be equidistant between the points.

9. Method for determining a carbon footprint of a device according to one of claims 1 to 8, in which before carrying out the step of incrementing (37) said carbon footprint (Ec) with a carbon cost per distance per country (C(ES) / km, C(FR)km) multiplied by the distance of the section in each country (D(Nl-Fl), D(F1-N2)), the method implements a step aimed at detecting whether a network link specific is used for the section and, if so, the incrementation step (37) of said carbon footprint (Ec) is carried out according to the expected cost of the specific network link detected.

10. Method for determining a carbon footprint of a device according to one of claims 1 to 9, in which the step of incrementing (37) said carbon footprint (Ec) with a carbon cost per distance per country (C(ES) / km, C(FR)km) multiplied by the distance of the section in each country (D(Nl-Fl), D(F1-N2)) also records the carbon cost of the section (R1-R5) in the database (19).

11. A method for determining a carbon footprint of a device according to claim 10, wherein the carbon cost of the section (R1-R5) is recorded in the database (19) with an expiry period so that, when this expiry period is reached, the carbon cost is deleted.

12. Method for determining a carbon footprint of a structure (10) comprising a set of peripherals (14), the carbon footprint of the structure (10) being determined by accumulating the carbon footprint (Ec) of the set of peripherals (14) according to one of claims 1 to 11.

Images