JP5117556B2 - Communication device - Google Patents

Description

  Embodiments described herein relate generally to a communication device that generates information related to a communication event based on a communication event that has occurred with another device.

  Various devices having specific functions have been distributed over a network, and the construction or use of distributed systems that function respectively has started (smart grids, distributed building management systems, etc.). In such a distributed system, devices perform necessary information communication through a network when necessary. Normally, devices involved in the communication generate certain information about the event based on a communication event that occurs between the devices and leave it as a log.

  On the other hand, in such a distributed system, it is generally difficult to grasp the relationship with the entire system only by monitoring individual devices. For example, when a failure occurs, tracking such as where to investigate in the system retroactively from the device where the failure was observed and how to take countermeasures is very important as the scale of the system increases. Complex and extremely difficult.

JP 2006-246065 A

  The present invention provides a communication apparatus capable of generating information related to a communication event based on a communication event that occurs between devices belonging to a distributed system so as to be suitable for efficient collection and search under a certain method. The purpose is to do.

  The communication apparatus according to the embodiment includes a communication control unit, an event encoding unit, and a transmission unit. The communication control unit specifies an attribute C of the communication event including at least the device identifier p of the other party of the communication event based on a communication event that has occurred with another device. The event encoding unit uses the given secure hash function h () and the given constant a from the given own device identifier n and the occurrence time t of the communication event to obtain the key value K as an expression K = h (n) + a · t, and a value corresponding to a set of n, t, C on a one-to-one basis using the device identifier n, the generation time t, and the attribute C of the communication event The value V is obtained. The sending unit sends the set of the key value K and the value value V to the external storage device.

The apparatus related figure explaining the relationship by which the communication apparatus of embodiment is arrange | positioned and connected. The block diagram which shows the main internal structures of the communication apparatus of embodiment. 2 is a conceptual diagram illustrating a data structure in a storage apparatus (distributed table) that is a transmission destination of the communication apparatus according to the embodiment. FIG.

  Based on the above, the communication device of the embodiment will be described below with reference to the drawings. First, FIG. 1 illustrates a relationship in which the communication apparatuses of the embodiment are arranged and connected.

  The communication devices N1, N2, N3, N4, (...) are node devices connected to this network so that they can communicate with each other via the network. For example, when a smart grid or microgrid is assumed as the entire network, the communication device N1 and the like include, for example, a controller of a generator, a controller mounted on a power converter such as a solar battery or wind power generation, an electric vehicle, or a plug-in. Hybrid vehicle charging device, home energy management system control device, commercial building energy management system control device, industrial equipment related control device, power system monitoring device, watt-hour meter (smart meter) with communication function, It is a communication device mounted on them as part of an automatic meter reading device that supports smart meters.

  These communication devices N1 and the like perform necessary information communication between the devices via a network (for example, the Internet) using a protocol corresponding to the function of the device. Normally, devices involved in the communication generate certain information about the event based on a communication event that occurs between the devices, and leave it as a log.

  FIG. 1 shows the occurrence of the following communication event as an example. That is, a communication event (subscribe) occurs from the communication device N1 to the same N2 at the time t1, another communication event (subscribe) occurs from the communication device N1 to the same N3 at the time t2, and the time Another communication event (unsubscribe) has occurred at t5. Further, another communication event (subscribe) occurs from the communication device N2 to the same N4 at the time t3, and another communication event (subscribe) occurs from the communication device N3 to the same N4 at the time t4. Yes.

  Thereby, in the communication device N1, “N1 @ t1 | SdN2”, “N1 @ t2 | SdN3”, and “N1 @ t5 | uSdN3” are generated and stored as logs. Logs are also generated and stored for the communications N2, N3, and N4 as shown in FIG. Here, the left side of the vertical bar (| symbol) of the log is an element related to both K (described later) and V (described later), and the right side is an element mainly related to V only. Further, “S” in the log is Subscribe (subscribe), “Sd” is Subscribe (subscribe: an abbreviation for a passive type of subscribe), “uS” is Unsubscribe, “USd” is Unsubscribed (an unsubscribed passive abbreviation for unsubscribed). In this example, the communication device N1 is a source of information, and information flows to the communication device N4 via the communication devices N2 and N3. In this case, the communication device N4 does not recognize that the communication device N1 is the transmission source.

  This embodiment allows information relating to communication events to be easily explored retrospectively from a communication device that is downstream in the information flow, in particular information relating to communication events to suit it. It is intended to be recorded in the storage device (distributed table) 100. Here, it is assumed that the storage apparatus 100 is a table that can write in the format of put (key, value) and can read in the format of get (key) → value. A distributed hash table can be used as the distributed table, but another method that satisfies the following conditions may be used.

  That is, in this distribution table, the data entry value can be registered in the key key with put (key, value), and then the data entry value is obtained from the key key by using get (key) → value. It is only necessary that the data entry value list [value] can be obtained from the key range [key1, key2) by using getRange (key1, key2) → [value].

  FIG. 2 shows the main internal configuration of each of the communication devices N1, N2, N3, N4,... (Hereinafter referred to as “communication device N1”). As illustrated in FIG. 2, the communication device N1 includes a communication control unit 11, an output management memory 12, an input management memory 13, a clock 14, an event encoder 15, a transmission unit 16, and a distributed table management unit 17.

  The communication control unit 11 is connected to another device (communication device) via, for example, the Internet, and includes at least a device identifier p of the other party of the communication event based on a communication event that occurs with the device. The attribute C of the event is specified. As an example of the attribute C, in addition to the device identifier p of the other party, a pair of communication event information flow direction d and communication event communication identifier (session identifier) s can be cited. The attribute C may include the occurrence time t of the communication event. The communication identifier s is specified as the same as that at the other party by exchanging with the other party. When a plurality of information flows with the same opponent occurs, another communication identifier s is exchanged and specified each time.

  The information flow direction d is the direction of information flow between itself and the other party involved in the communication event. The direction d is defined as a substantial information flow even if one is positioned as a server and the other as a client in the communication event, and both are defined as possible sources of information. For example, when one is a meter-reading server and the other is a smart meter (client), the client transmits information to the meter-reading server using the smart meter. The same applies when the smart meter falls into an abnormal state and notifies the meter-reading server. Conversely, when a load control command is sent from the meter reading server to the smart meter, the meter reading server is the information sending side.

  Further, as an example of the attribute C different from the above, the following may be used instead of the direction d of the communication event information flow. That is, instead of d, this communication event occurred when subscribed to the other party of the communication event, occurred when subscribed from the other party of the communication event, or occurred when unsubscribed to the other party of the communication event. Information e indicating which one of the four types is generated by being unsubscribed from the other party of the communication event or the communication event. This case can be suitably applied when the network to which the communication device N1 participates is a so-called publish-subscribe network.

  The communication control unit 11 records and saves a log including the attribute C in the output management memory 12 in the case of a communication event in which the communication control unit 11 is the information source. In the case of a communication event for which the communication control unit 11 has received information, the communication control unit 11 records and saves a log including the attribute C in the input management memory 13. These are for communication event management in each communication device (communication device N1, etc.). The output management memory 12 and the input management memory 13 are memories that can store such log information. The communication control unit 11 also outputs the identified attribute C to the event encoder 15. This output includes information on time t.

  The clock 14 is a clock that provides time information in order for the communication control unit 11 to specify the occurrence time t of the communication event that has occurred. As the time t, UNIX (registered trademark) standard time or the number of seconds elapsed from a certain common reference point can be used.

  The event encoder 15 uses the given secure hash function h () and the given constant a from the given device identifier n and the communication event occurrence time t passed from the communication control unit 11. The key value K is obtained according to the equation K = h (n) + a · t. Further, by using the own device identifier n and the occurrence time t and the communication event attribute C passed from the communication control unit 11, a value value V corresponding to one-to-one is obtained for a set of n, t, and C. .

  For example, in the case of an IP network, the own apparatus identifier n is obtained by passing a combination of an IPv4 address or an IPv6 address, a transport layer protocol (such as TCP / UDP), and a port number through a secure hash function. Can be used. Alternatively, a UUID (universally unique identifier) generated from time or a random number can be used instead. Although these circumstances will be described before and after, the same applies to the device identifier p of the other party of the communication event specified by the communication control unit 11 in the other device.

  As the secure hash function h () given to the event encoder 15, a known secure hash function h (k) → {x | xεH} (where H is a hash value space) such as SHA-1 is used. it can. The constant a is a constant that defines how much the time advances the key value K, which is a hash value, in the hash value space H. In other words, the constant a can be said to be a parameter for appropriately allocating information sent from the specific communication device (N1,...) To the storage device 100 in the distribution table.

For example, the constant a can be determined as follows. As a premise, t is expressed in seconds. The bit width b of the hash value space H is determined by the secure hash function h (k), but assuming that b = 128 bits, the value value V so that the hash value space H goes around once in one year (31536000 seconds). Is assigned to a = 2 ¹²⁸ /31536000=1.08×10 ³¹ . In short, as the constant a, the number of powers of 2 may be a = 2 ¹⁰³ (= 10 ^31.0061 ) in this case. When a power of 2 is used as the constant a, the operation a · t is a bit shift in the equation K = h (n) + a · t, and is simplified.

  In order to obtain a value value V corresponding to a set of n, t, and C on a one-to-one basis, for example, using a given list function list (), V = list (t, d, n, p, s (If e is used instead of d, V = list (t, e, n, p, s) is used). Incidentally, when such a listing function list () is used, if an unlisting function unlist () that is an inverse function is used, t, d, n, p can be obtained by unlist (V) using V as an argument. , S (or t, e, n, p, s).

  The key value K and the value value V obtained by the event encoding unit 15 are sent from the event encoding unit 15 to the sending unit 16. The sending unit 16 sends and registers the sent K and V pairs to the storage apparatus 100. Even when the storage apparatus 100 uses a distributed hash table, it is not necessary to apply a secure hashing function to K in this table. This is because the event encoding unit 15 has already performed the calculation of h (n) and has obtained the same effect, and to facilitate the retrieval of information related to continuous time-series communication events.

  The distributed table management unit 17 is not directly related to the operation described above, but has the following functions. That is, the status of the storage apparatus 100 is periodically checked to check available distributed table nodes. If the distributed table used by the storage apparatus 100 is a distributed hash table and uses, for example, the Chord algorithm, a search with a random value set to K is performed, and table nodes that can be used by recursive queries are found. Learn.

  Next, FIG. 3 conceptually illustrates a data structure in a storage apparatus (distributed table) that is a transmission destination of the communication apparatus of the embodiment. In the figure, symbols representing communication events are the same as those in FIG.

  In a distributed table such as a distributed hash table, it is an essential requirement for efficient operation that data writing and access are distributed. Normally, by applying a secure hash function to an arbitrary key, it is possible to distribute access with a uniform hash value distribution for a group of keys having an arbitrary distribution.

  In the embodiment described above, various devices in the world may communicate with each other. In this case, the diversity of devices that perform communication is smaller than the diversity of communication itself. Therefore, if the local device identifier n is used as a key, hashed with h (n) and written in the distributed hash table, the amount of data that must be written in a specific table node increases linearly over time. . On the other hand, if the communication itself is hashed with the key and written in the distributed hash table, the communication event has a meaning in time series, and the processing becomes very difficult when trying to use (search) it. End up.

  Therefore, in the embodiment described above, information on communication events in the vicinity of the time is stored in a continuous area of the table node and the load is distributed over time while ensuring dispersibility for each device by h (n). As is done, the key is h (n) plus a · t.

  In this way, for example, the occurrence of a communication event in the communication devices N1 and N2 is conceptually stored in a table as shown in the upper part of FIG. That is, the starting point of the storage location is h (n), which is determined for each communication device, so that it seems that a distributed table exists for each communication device in a pseudo manner. Actually, all the communication devices share one distribution table (the lower part of FIG. 3). The use of the storage device (distributed table) 100 will be further described in the application example described below.

(Application example 1)
Consider a smart grid as a distributed system. An electric vehicle charging device is connected to the smart grid. A controller mounted on the solar cell is also connected. Many devices other than these are connected to the smart grid, and each time a communication event occurs between these devices, each device sends a key value to the storage device (distributed table) 100 as described above. A set of K and value value V is transmitted and registered. Here, it is assumed that the power supply to the charging device is stopped.

  Here, for example, it is assumed that the steps leading to the stop of power supply are as follows. 1. Decrease in power generated by local solar cells. 2. It is predicted by the local EMS (energy management system) that the power supply amount of the entire region becomes unstable. 3. Supply / demand adjustment requests to local charging device operators. 4). Power supply stop instruction to the charging device from the charging device operating company.

At this time, from the side of the charging device that observed only the result of the fourth step, the cause of the power supply stop is not clear. Therefore, the EMS computer mounted on the charging device (or electric vehicle) acquires the self device identifier n ₀ of the charging device, and K = h (n ₀ ) + a from this n ₀ and the event occurrence time t _0. - to calculate the _{t 0.} Then, getRange (K−td, K + td) is executed on the storage apparatus (distributed table) 100 in consideration of the assumed error width td of a · t ₀ . In setting the error width td, for example, a value obtained by calculating a delay time due to communication + a time shift coming from a clock shift + an expected delay time from the reception of the command to the occurrence of an event. Consider.

The search described above to the storage device (the dispersion table) 100, existing near a time t _0, a communication event associated with the charging device, can be obtained in the form of a list of V [V]. Then, using each value value V as an argument, for example, by executing unlist (V), a set of t, d, n, p, s (or a set of t, e, n, p, s) is obtained. Can be sought. Since these pieces of information include the device identifier p of the other party of the communication event, in the case of this example, it can be understood that the device of the charging device operating company may be checked by the device identifier p.

Therefore, next, the device identifier p of the charging device operating company is newly set as its own device identifier n- _1, and the issue time of the stop command is set to t- ₁ , and K- ₁ = h (n- ₁ ) + a.t. ₋₁ is calculated, and getRange (K ₋₁ -td, K ₋₁ + td) is executed for the storage apparatus (distributed table) 100. Thereafter, it can be traced back to the device in the regional EMS in the same manner as described above.

  Thereafter, similarly, the EMS computer mounted on the charging device (or electric vehicle) searches the storage device (distributed table) 100 to determine the cause of the event up to the controller mounted on the source solar cell. You can go back. This presents the user with a choice of whether to wait for the weather to recover, pay a special fee to buy another power, or stop charging at the current charge, for example. be able to.

(Application example 2)
Next, consider a distributed building management system as an example of a distributed system. In such a system, each communication device is mounted as a part of a sensor, an actuator, a building information management system, or the like provided in each building.

  In distributed building management systems, in general, by comparing and examining the operating environments of multiple buildings, it is possible to predict failures from phenomena such as sudden increase in power consumption even for the same type of equipment, and to save energy from differences in usage patterns. Development such as integrated building management or ESCO business, such as creating action guidelines, can be considered. As a result, a system is built that goes beyond simple building management.

  In such a distributed building management system, there is a problem that it becomes difficult to see which tenant's data the sensed data is or what data the output of the integrated building management system is based on. Also, what statistics are derived from which tenants when building management companies with two tenants that are derived from this and should not originally mix data (eg contractual reasons) create integrated data? There is also a problem that there is room for human error to occur because it takes time and effort to investigate.

  In contrast, when the communication device described above is installed in each device of the distributed building management system, and the communication device described above is also installed in a calculation device (calculation node, calculation process) for processing intermediate integrated data. There are the following advantages. That is, by searching the storage apparatus 100 in which communication events that occur between these apparatuses are registered, it is easy to clarify which data is connected to which result via which processing routine. can do.

  For example, when there are tenants A and B that are mutually exclusive, if the average power consumption data of the two parties and the individual power consumption data are clear, the power consumption of the other tenant can be known. I can do it. This is not desirable because the activity of the business activity can be estimated from the power consumption data.

  Therefore, it is assumed that the integrated building management company sets a rule so that the exclusive tenant A and the tenant B do not give information including each other's data to the exclusive tenant A and tenant B, respectively. In such a case, this trader searches for the storage device 100 using the data exit device, and inspects the communication event registered by the communication device installed in the sensor or process by going back in time. It is easy to check whether aggregated values that contain the data are output.

  In addition, in order to equip the above-described communication device equivalently to a calculation device (calculation node, calculation process) that processes intermediate integrated data, for example, an intervention in a library (such as a dll hook) during communication is performed. The technique used can be adopted. In this case, in the registration to the storage apparatus 100 and the search for the storage apparatus 100, a well-known data history technique can be applied to improve convenience.

  As described above, according to the communication device of the embodiment, this information can be generated so that information related to communication events between devices can be efficiently collected. As a result, the causal relationship between communication events can be easily checked, and dynamic communication event information management for a large number of communication devices can be performed. It is also possible to easily build information traceability applications on distributed systems.

  As mentioned above, although several embodiment of this invention was described, these embodiment is shown as an example and is not intending limiting the range of invention. These novel embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the spirit of the invention. These embodiments and modifications thereof are included in the scope and gist of the invention, and are included in the invention described in the claims and the equivalents thereof.

  DESCRIPTION OF SYMBOLS 11 ... Communication control part, 12 ... Output management memory, 13 ... Input management memory, 14 ... Clock, 15 ... Event encoder, 16 ... Sending part, 17 ... Distributed table management part, N1, N2, N3, N4 ... Communication apparatus ( Node), 100... Storage device (distributed table).

Claims (4)

A communication control unit that identifies an attribute C of the communication event, including at least a device identifier p of the other party of the communication event, based on a communication event that has occurred with another device;
From the given device identifier n and the occurrence time t of the communication event, using the given secure hash function h () and the given constant a, the key value K is expressed by the equation K = h (n) + a An event for obtaining a value value V corresponding to a set of n, t, C on a one-to-one basis using the own device identifier n, the occurrence time t, and the attribute C of the communication event. An encoding section;
A communication device comprising: a sending unit that sends the set of the key value K and the value value V to an external storage device. The attribute C of the communication event specified by the communication control unit includes the communication event information flow direction d, the communication event communication identifier s, in addition to the device identifier p of the communication event counterpart. Have
The apparatus according to claim 1, wherein the value value V obtained by the event encoding unit is obtained by V = list (t, d, n, p, s) using a given listing function list (). The attribute C of the communication event specified by the communication control unit is generated in addition to the device identifier p of the other party of the communication event, and the communication event is generated by subscribing to the other party of the communication event; Whether the event occurred when subscribed from the other party of the communication event, occurred when unsubscribed to the other party of the communication event, or occurred when unsubscribed from the other party of the communication event And the communication identifier s of the communication event,
The apparatus according to claim 1, wherein the value value V obtained by the event encoding unit is obtained by V = list (t, e, n, p, s) using a given listing function list ().   The apparatus according to claim 1, wherein the constant a used by the event encoding unit is a power of two.

Images