JP2020071691A - Software build process recording system and method, and software build process monitoring program - Google Patents

Abstract

To provide a system suitable for recording a software build process.SOLUTION: A system (100) includes: a node (10) having a build process monitoring function (11) for generating build process data including a command line, a command binary file, a read file, and hash values of the generated files and sending the build process data to a network (20); a node (10) having a block generation function (12) for generating a block which includes a block header containing a hash value, a time stamp and a nonce value of a block header of the currently latest block, and a top hash of plural pieces of unregistered build process data, and plural pieces of build process data, and sending the block to the network; and a node (10) having a block holding function (13) for retaining blocks that have been generated so far, and newly retaining the block received from the node having the block generation function.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a system and method for recording a software build process, and a program for causing a computer to monitor the software build process.

  Software traceability, which makes it possible to verify the development process of software products afterwards, is one of the important qualities of software products. According to the definition of CoEST (Center of Excellence for Software & Systems Traceability), software traceability means "to associate uniquely identifiable artifacts that make up a software product with each other, maintain necessary links, and connect those networks. The ability to use and answer questions for both the software product and its development process. " Such software traceability can be strong evidence in litigation dispute resolution. Therefore, software traceability is particularly important in areas where defects existing in software can pose a great threat to human life and the environment.

  As information that ensures software traceability, it is costly and investment effect is low to manually create links between artifacts such as requirement specifications, design documents, and source codes that are accumulated in large amounts inside companies and organizations. It has been pointed out. Therefore, a technique has been developed for automatically generating a related link from a group of information accumulated in a company or an organization (for example, Non-Patent Documents 1 and 2). Also, in software development using open source software (OSS) components, in order to evaluate the license compliance of each component, system calls that occur at the time of building the software are monitored to automatically build the dependency relationships of each component (for example, , Non-Patent Document 3).

Hazeline U. Asuncion, Arthur U. Asuncion, Richard N. Taylor, "Software traceability with topic modeling," Proceedings of the 32nd ACM / IEEE International Conference on Software Engineering, pp. 95-104, May 01-08, 2010. N. Ali, Y.-G. Gueheneuc, G. Antoniol, "Trustrace: Mining software repositories to improve the accuracy of requirement traceability links", IEEE Trans. Software Eng., Vol. 39, no. 5, pp. 725- 741, 2013. S. van der Burg, E. Dolstra, S. McIntosh, J. Davies, DM German, A. Hemel, "Tracing software build processes to uncover license compliance inconsistencies," Proc. Of the 29th ACM / IEEE Intl. Conf. On Automated Software Engineering (ASE), pp. 731-742, 2014.

  In recent years, due to the globalization of software development, the number of geographically dispersed development bases participating in one development project is increasing. For example, a company based in North America may outsource software development to a company based in Asia. Since the decentralization of development bases is expected to improve development speed and reduce costs, the trend of globalization and decentralization of software development is expected to continue to expand. On the other hand, it becomes a problem that timely communication is insufficient due to the influence of the time difference, and it becomes difficult to unify the directionality of the parties concerned and to build a trusting relationship. In particular, when utilizing information that guarantees software traceability among multiple companies or organizations, standardization and sharing methods of such information remain unsolved or partially solved.

  On the other hand, in recent years, blockchain has begun to be utilized as an approach to traceability assurance in fields other than software, such as medical care and food. However, it is difficult to apply the traceability mechanism used in the fields such as medical care and food directly to the field of software development. Because, 1) software products consist of hundreds to tens of thousands or more of artifacts, and 2) artifacts (especially source code) that make up software products have a high risk of being copied or diverted. Because there is.

  In view of the above problems, it is an object of the present invention to provide a system suitable for recording a software build process, and a program used in such a system for causing a computer to monitor the software build process.

  According to one aspect of the present invention, there is provided a system for recording a software build process in a block chain on a network formed by a plurality of computers, the command line being executed in the software build process, and the command in the command line. A build process including a binary file, a file specified by the command line, a file other than the specified file that is called when the command line is executed, and a hash value of each file obtained by executing the command line A plurality of nodes having a build process monitoring function for generating data and transmitting the data to the network, a hash value of a block header of a current latest block, a time stamp, a nonce value selected so as to satisfy a predetermined condition, and a previous A block header including top hashes of a plurality of build process data not registered in the block chain received from a plurality of nodes having a build process monitoring function, and a block including the plurality of build process data are generated and transmitted to the network. A node having a block generating function, and a plurality of nodes having a block retaining function that retains the blocks generated so far and newly retains the block received from the node having the block generating function. A software build process recording system is provided.

  According to another aspect of the present invention, there is provided a program for causing a computer to monitor a software build process, the command line obtaining unit obtaining a command line input to the computer, and the command line obtaining unit. Specify each path of the command binary file, the file specified in the command line, the file other than the specified file that is called when the command line is executed, and the file obtained by executing the command line File path identifying means, hash value computing means for computing each hash value of the file existing in each of the obtained command line and each of the identified paths, process including the obtained command line and each of the identified paths Generate summary And a plurality of build process data generating means for storing the obtained command line and each specified path in the process summary into the calculated hash value to generate build process data, There is provided a software build process monitoring program that causes a computer to function as a transmission unit that transmits the generated build process data to a network formed by the computer.

  According to the present invention, by recording a software build process using a blockchain, it is possible to record a large number of software build processes globally without geographical and organizational restrictions. Verification is also easy. In addition, only the hash value of the command line and the hash value of the file related to the execution of the command line are written in the build process data, and it is not necessary to disclose the artifacts (particularly the source code) that make up the software product. The confidentiality is maintained.

Overall configuration diagram of a software build process recording system according to an embodiment of the present invention Block diagram of software that realizes a build process monitoring function according to an example Build process data flow chart Schematic diagram of build process data generation example when the compile command is executed Schematic diagram of blockchain used in software build process recording system Diagram explaining the post-verification process of the build process

  Hereinafter, embodiments will be described in detail with reference to the drawings as appropriate. However, more detailed description than necessary may be omitted. For example, detailed description of well-known matters and duplicate description of substantially the same configuration may be omitted. This is to avoid unnecessary redundancy in the following description and to facilitate understanding by those skilled in the art.

  It is to be noted that the inventor provides the accompanying drawings and the following description in order for those skilled in the art to fully understand the present invention, and is not intended to limit the subject matter described in the claims by these. Absent.

≪System configuration example≫
FIG. 1 is an overall configuration diagram of a software build process recording system according to an embodiment of the present invention. A software build process recording system (hereinafter referred to as “process recording system”) 100 according to the present embodiment includes a plurality of nodes (computers such as personal computers and workstations) 10 distributed and arranged on the Internet. Although eight nodes 10 are illustrated in FIG. 1 for the sake of convenience, the number of nodes 10 included in the process recording system 100 is not limited to eight, and a small-scale configuration in which the number of nodes 10 is less than eight, or A large scale configuration in which tens or hundreds of nodes 10 participate is also possible.

  There are no geographical or organizational restrictions on the node 10, and various software vendors and software house computers around the world and end user computers around the world can be the node 10. For example, a computer belonging to the A company in North America, a computer belonging to the B company in Asia, a computer belonging to the end user C in Japan, etc. can be components of the process recording system 100.

  In the process recording system 100, the nodes 10 form a peer-to-peer type network 20. That is, in the process recording system 100, there is no hierarchical relationship such as client-server between the nodes 10, and the nodes 10 communicate with each other on an equal footing. Note that each node 10 is not necessarily directly connected to all the other nodes 10, and there may be a node 10 that communicates with the partner node 10 via another node 10.

  The information of the software build process that should be left as evidence of software development in the process recording system 100 having the above configuration is collected in blocks (blocks 50 in FIG. 5) that are blocks of data at predetermined time intervals. A chain, that is, a block chain is distributed and held by a plurality of nodes 10. In this way, the process recording system 100 records the information of the software build process by applying the block chain technology which is one of the distributed ledgers.

  The process recording system 100 has three functions of a build process monitoring function 11, a block generation function 12, and a block holding function 13 in order to record information on a software build process using a block chain. These three functions are independent of each other, and any one or a plurality of functions can be installed in one node 10. Of these, if there is at least one node 10 having the block generation function 12, the process recording system 100 functions effectively as a whole. Further, in order to make it possible to verify any build process afterwards, it is sufficient that the process recording system 100 has at least one node 10 having the build process verification function 14. Hereinafter, each function will be described in order.

≪Build process monitoring function≫
FIG. 2 is a block diagram of software that implements a build process monitoring function according to an example. FIG. 3 is a flowchart of build process data generation processing in the node 10 having a build process monitoring function. FIG. 4 is a schematic diagram of an example of build process data generation when a compile command is executed. Hereinafter, the build process monitoring function will be described in detail with reference to these drawings.

  The build process monitoring function 11 is a function that is mainly installed in a computer on the software developer side such as a software house or a software vendor, and monitors software development work performed using the computer, and the executed command line and its Information about input / output files related to execution is converted to build process data in a predetermined format and transmitted to the network 20.

  As shown in FIG. 2, the build process monitoring function 11 can be realized as a kernel of the operating system (OS) 101 or as middleware interposed between the OS and the application program (AP) 102. Specifically, the build process monitoring function 11 includes a command line acquisition unit 111, a file path identification unit 112, a hash value calculation unit 113, a build process data generation unit 114, and a transmission unit 115.

  When a software developer develops software, a command is directly input from a keyboard to a CUI (Character User Interface) such as a console or a terminal, or a command is input through a GUI (Graphical User Interface) provided by a software development application. Compile, build, make, etc. are executed, and source code files, intermediate binary files, library files, etc. are read to generate object files. The command line acquisition unit 111 monitors commands such as compile, build, and make that the software developer inputs to the computer through the AP 102 or the console of the OS 101, and acquires the input command line (S1 in FIG. 3). In the example of FIG. 4, the command line acquisition unit 111 recognizes that the compile command “gcc” has been input to the console and includes “LANG = C / usr / bin / gcc /mysrc.c -o myprogram I got the command line ".

  The file path identification unit 112 identifies each path of the file related to the execution of the command line acquired by the command line acquisition unit 111 (S2 in FIG. 3). Specifically, the file path identification unit 112 appropriately supplements the file path when the file path is omitted, and a command binary file (hereinafter, referred to as “command binary”) or a file specified on the command line (for example, Paths of source code, Makefile, etc.) and other files that are called when the command line is executed (hereinafter referred to as "read file") and files obtained by executing the command line (hereinafter referred to as "generated file") Identify. In the example of FIG. 4, the file identification unit 112 has a command binary file path of “/ usr / bin / gcc” and read file file paths of “/mysrc.c” and “/ usr / include / stdlib. h ”,“ /usr/lib/glibc.o ”, and specifies that the file path of the generated file is“ / myprogram ”. In this way, the file path identification unit 112 also identifies paths such as header files and library objects listed in the source code “/mysrc.c” that are not specified in the command line.

  The hash value calculation unit 113 calculates each hash value of the command line acquired by the command line acquisition unit 111 and the command binary, the read file, and the generated file existing in each path identified by the file path identification unit 112 (Fig. 3 S3). Specifically, the hash value calculation unit 113 uses hash functions such as RIPEMD (RACE Integrity Primitives Evaluation Message Digest), RIPEMD-160, MD5 (Message Digest Algorithm 5), SHA (Secure Hash Algorithm) -256, and SHA-3. Is used to receive an arbitrary character string of an arbitrary length or arbitrary data of an arbitrary size, and generate an output of a predetermined size.

  The build process data generation unit 114 executes the process summary 30 including the command line acquired by the command line acquisition unit 111 and the command binary, read file, and generated file path existing in each path specified by the file path specifying unit 112. It is generated and stored in a storage device (not shown) (S4 in FIG. 3). Further, the build process data generation unit 114 replaces the command line and the path of each file in the process summary 30 with each hash value calculated by the hash value calculation unit 113 to generate the build process data 40 (S5 in FIG. 3). ).

  Further, the build process data generation unit 114 includes, in the process summary 30, documents such as software metrics, requirement specifications, and software installation procedures, as well as paths of all types of files related to software build such as audio, images, and videos. The build process data 40 can be generated by replacing the paths with the hash value calculated by the hash value calculation unit 113 (hash value of the file existing in each path).

  In this way, the process summary 30 shows what command line is executed in the software build process, which file is read and which file is generated, and what kind of file is associated. It represents file link information. On the other hand, the build process data 40 replaces the command line and the input / output file path described in the process summary 30 with the hash value (the file path is replaced with the hash value of the file existing in the path), and the build process concretely It is a concealment of what command line was executed, which file was read, and which file was created. Since the process summary 30 and the files existing in each path described in the process summary 30 are required later in the verification of the software build process, it is important that each company strictly stores them so as not to be deleted.

  The transmission unit 115 transmits the build process data 40 generated by the build process data generation unit 114 to the network 20 (S6 in FIG. 3). More specifically, the transmission unit 115 electronically signs the build process 40 to record when and by whom the build process data 40 was created, and then transmits the build process data 40 to the network 20. The build process data 40 may be broadcast to the network 20 or may be unicast or multicast to the specific node 10 having the block generation function 12.

≪Block generation function≫
FIG. 5 is a schematic diagram of a block chain used in the process recording system 10. The build process data 40 transmitted to the network 20 in the process recording system 100 is collected into blocks 50 at predetermined time intervals, and each block 50 is connected in such a manner as to hold the hash value of the immediately preceding block 50 to form the block chain 200. It is formed. The block 50 includes a block header 51 and a plurality of build process data 40. The block header 51 includes a hash value 52 of the block header 51 of the immediately preceding block 50, a time stamp 53 indicating the generation time of the block 50, a nonce value 54, and a plurality of build process data 40 included in the block 50. Including the top hash 55.

  The node 10 having the block generation function 12 is transmitted from the node 10 having the build process monitoring function 11 to the network 20 and collects unregistered build process data 40 in the block chain 200 to generate a new block 50. For example, when the block number of the current latest block 50 is [n] and the block 50 of the block number [n + 1] is generated, the node 10 having the block generation function 12 determines the block of the block 50 of the block number [n]. The hash value 52 of the header 51 is calculated, the unregistered build process data 40 is collected in the block chain 200, the top hash 55 corresponding to the root of the Merkle tree is calculated, and the time stamp 53 and an appropriate nonce value 54 are calculated. Set and generate the block header 51. Then, the node 10 having the block generation function 12 repeatedly changes the nonce value 54 until a nonce value 54 is found such that the hash value of the generated block header 51 becomes smaller than the target value, and the nonce value 54 satisfying the condition is satisfied. If found, the block header 51 is confirmed and the block 50 having the block number [n + 1] is transmitted to the network 20. The block 50 may be broadcast to the network 20 or may be multicast to the plurality of nodes 10 having the block holding function 13.

  In the process recording system 10, any type of public, private, and consortium can be applied as the block chain 200. For example, when the public block chain 200 is applied, anyone can freely participate in the process recording system 100 as the node 10 having the block generation function 12. In this case, the generation function of the block 50 can be maintained by introducing a consensus algorithm such as PoW (Proof of Work) that pays the reward to the person who generated the block 50 earliest. On the other hand, when the private or consortium type blockchain 200 is applied, an administrator is set in the process recording system 100, and only the administrator or a person authorized by the administrator records the process as the node 10 having the block generation function 12. You can participate in the system 100.

≪Block holding function≫
The node 10 having the block holding function 13 holds the block 50 generated so far, and newly holds the block 50 received from the node 10 having the block generating function 12. In the case of the private type or consortium type blockchain 200, since the block 50 generated by the node 10 having the block generation function 12 is reliable, the node 10 having the block holding function 13 may hold it as it is. In the case of the type block chain 200, it is necessary to verify the correctness of the block 50 generated by the node 10 having the block generation function 12. Therefore, the node 10 having the block holding function 13 verifies the correctness of the block 50 received from the node 10 having the block generation function 12, and newly holds the block 50 if the verification is successful. Good. More specifically, the node 10 having such a block holding function 13 determines whether the hash value 52 of the previous block included in the block header 51 of the block 50 received from the node 10 having the block generation function 12 is correct. Whether or not the nonce value 54 satisfies a predetermined condition, a top hash of the plurality of build process data 40 included in the block 50 is calculated, and the calculated top hash 55 is included in the block header 51 of the block 50. The correctness of the block 50 is determined by verifying whether they match. Only the blocks 50 correctly generated in this way are added to the block chain 200.

≪Build process verification function≫
In the process recording system 100, if the software is not loaded with a virus or spyware, or if there is a suspicion about the reported software metrics, anyone can use the build process verification function 14 to obtain the desired build process data 40. Can be verified after the fact. In the post-verification of the build process data 40, the process summary 30 corresponding to the build process data 40 and all the files listed in the process summary 30 are obtained from the company or organization involved in the generation of the build process data 40. There is a need to.

  FIG. 6 is a diagram illustrating the post-verification processing of the build process. The node 10 having the build process verification function 14 acquires the arbitrary build process data 40 included in the arbitrary block 50 held by the arbitrary node 10 having the block holding function 13. On the other hand, the node 10 having the build process verification function 14 reproduces the build process data 40A from the process summary 30 and the file 31 separately obtained from the company or organization. This reproduction procedure is the same as the build process data generation described above. Then, the node 10 having the build process verification function 14 compares the acquired build process data 40 with the reproduced build process data 40A, determines that the build process data 40 has not been tampered with if both match, and matches. If not, it is determined that there is tampering.

<< Effect >>
According to this embodiment, the following effects are achieved.
1) A large number of software build process records can be carried out globally without geographical and organizational restrictions.
2) The newly generated build process data 40 is collected and connected to the block 50 one after another to ensure the order relation of the software build.
3) Only the hash value of the command line and the hash value of the file related to the execution of the command line are written in the build process data 40, and it is necessary to disclose the artifacts (particularly the source code) that constitute the software product. The confidentiality is maintained.
4) By sharing the build process data 40 with many participants using the blockchain 200, it becomes difficult to tamper with the build process. In addition, anyone can verify any build process afterwards, ensuring software traceability.
5) When the public blockchain 200 is applied, an administrator of the process recording system 100 is not required, and a platform that anyone can freely participate can be realized.
6) By installing the build process monitoring function 11 in the computer, the computer automatically generates the build process data 40 and transmits it to the network 20 in the back of software development work, and the build process data 40 is recorded. Can develop software without being aware of the build process records.

≪Modification≫
In addition to generating the build process data 40 at the timing when the command is executed, even if the command is not executed, a snapshot of the target file may be taken at a predetermined timing and the hash value may be registered in the build process data 40. Good. For example, it is effective for files handled by continuously operating systems such as databases.

  In addition, since it is not necessary to record the build process of the temporary test file that is not related to the final software product in the process recording system 100, it is not necessary to enable / disable the build process monitoring function 11. It is also possible not to generate the appropriate build process data 40.

  Instead of generating the build process data 40 every time one command line is executed, one build process data 40 may be generated by collectively executing a plurality of command lines. Thereby, the transmission frequency of the build process data 40 can be reduced.

  As described above, the embodiment has been described as an example of the technique of the present invention. To that end, the accompanying drawings and detailed description are provided.

  Therefore, among the constituent elements described in the accompanying drawings and the detailed description, not only constituent elements essential for solving the problem but also constituent elements not essential for solving the problem in order to exemplify the above technology Can also be included. Therefore, it should not be immediately recognized that the non-essential components are essential, because the non-essential components are described in the accompanying drawings and the detailed description.

  Further, since the above-described embodiment is for exemplifying the technique of the present invention, various changes, replacements, additions, omissions, etc. can be made within the scope of the claims or the scope equivalent thereto.

  100 ... Software build process recording system, 10 ... Node (computer), 11 ... Build process monitoring function, 12 ... Block generation function, 13 ... Block holding function, 14 ... Build process verification function, 20 ... Network, 111 ... Command line acquisition Part, 112 ... File path acquisition unit, 113 ... Hash value calculation unit, 114 ... Build process data generation unit, 115 ... Transmission unit, 30 ... Process summary, 40 ... Build process data, 50 ... Block, 51 ... Block header, 52 ... hash value of the previous block, 53 ... time stamp, 54 ... nonce value, 55 ... top hash, 200 ... block chain

According to one aspect of the present invention, a system for recording a software build process in a block chain on a network formed by a plurality of computers , the hash value being an irreversible conversion of a command line executed in the software build process. A hash value that is an irreversible conversion of a binary file of a command on the command line , a hash value that is an irreversible conversion of a file specified on the command line, and a file that is called when the command line is executed and is not the specified file multiple having build process monitoring function of the hash value is irreversible conversion, and generate and build process data including a hash value is irreversible conversion of files obtained by executing the command line sent to the network And a node, irreversible conversion is a hash value of the block header of the current state block, timestamp, nonce value chosen for a given condition is satisfied, and received from a plurality of nodes having the build process monitoring function, block A block header including a top hash of a plurality of build process data not registered in the chain, a node having a block generation function of generating a block including the plurality of build process data and transmitting the block to the network, and a node generated so far. And a plurality of nodes having a block holding function for newly holding the block received from the node having the block generating function, and a software build process recording system.

According to another aspect of the present invention, there is provided a program for causing a computer to monitor a software build process, the command line obtaining unit obtaining a command line input to the computer, and the command line obtaining unit. Specify each path of the command binary file, the file specified in the command line, the file other than the specified file that is called when the command line is executed, and the file obtained by executing the command line File path specifying means, hash value that is an irreversible conversion of the acquired command line , and hash value calculation means that calculates a hash value that is an irreversible conversion of a file existing in each of the specified paths, and the acquired command line And said specified A process summary including a path is generated and stored in a predetermined storage area, and the obtained command line and the specified path in the process summary are replaced with the calculated hash values to generate build process data. There is provided a software build process monitoring program that causes a computer to function as a build process data generating unit for generating and a transmitting unit for transmitting the generated build process data to a network formed by a plurality of computers.

Claims (7)

A system for recording a software build process in a blockchain on a network formed by multiple computers,
A command line executed in the software build process, a binary file of the command in the command line, a file specified in the command line, a file other than the specified file called when the command line is executed, and the command A plurality of nodes having a build process monitoring function for generating build process data including each hash value of the file obtained by executing the line and transmitting the build process data to the network;
A hash value of the block header of the current latest block, a time stamp, a nonce value selected to satisfy a predetermined condition, and a plurality of unregistered blocks in the block chain received from the plurality of nodes having the build process monitoring function. A block header including a top hash of build process data, and a node having a block generation function for generating a block including the plurality of build process data and transmitting the block to the network,
A software build process recording system, comprising: a plurality of nodes having a block holding function, which holds the blocks generated so far, and newly holds the blocks received from the node having the block generating function.   Comparing any build process data included in any block held in any node having the block holding function with the build process data reproduced from the command line or file from which the build process data was generated The software build process recording system according to claim 1, further comprising a node having a build process verification function for verifying whether the command line and the file have been tampered with.   3. The software build process recording system according to claim 1, wherein the build process monitoring function generates build process data that further includes hash values of various files such as documents, sounds, images, and movies related to software build.   4. The software build according to claim 1, wherein the block holding function verifies the correctness of the block received from the node having the block generation function, and newly holds the block if the verification is successful. Process recording system. A method of recording a software build process on a blockchain on a network formed by a plurality of computers,
The specified file, in which a plurality of arbitrary nodes are executed in a software build process, a binary file of the command in the command line, a file specified in the command line, and a file called when the command line is executed. And a build process data including each hash value of the file obtained by executing the command line, and transmitting to the network,
An arbitrary node receives from a plurality of nodes having a hash value of a block header of the current latest block, a time stamp, a nonce value selected so as to meet a predetermined condition, and the build process monitoring function. Generating a block header including a top hash of a plurality of unregistered build process data and a block including the plurality of build process data and transmitting the block to the network;
A method of recording a software build process, comprising a step of newly holding a received block by any of a plurality of nodes holding the blocks generated so far.   Any node can output any build process data contained in any block held by any number of nodes and the build process data reproduced from the command line or file from which the build process data was generated. The software build process recording method according to claim 5, further comprising a step of comparing and verifying whether or not the command line and the file have been tampered with. A program that allows a computer to monitor the software build process,
Command line acquisition means for acquiring the command line input to the computer,
A binary file of the command in the acquired command line, a file specified in the command line, a file other than the specified file that is called when the command line is executed, and obtained by executing the command line File path identification means for identifying each path of the file,
Hash value calculation means for calculating each hash value of the file existing in the obtained command line and each of the identified paths,
A process summary including the acquired command line and the specified paths is generated and stored in a predetermined storage area, and the acquired command line and the specified paths in the process summary are calculated. Software build that causes a computer to function as a build process data generation unit that replaces each generated hash value to generate build process data, and a transmission unit that transmits the generated build process data to a network formed by a plurality of computers. Process monitoring program.

Images