Method and system for client/server and peer-to-peer disk imaging6144992Abstract A method and system for imaging data between two or more digital computers across a computer network is described, where the digital computers transfer data in a peer-to-peer mode and/or a client/server mode upon command of the operator. This invention address the problem of managing, updating and installing executable software, such as operating systems, utilities and application software packages on a large number of networked computer systems. By using this invention properly, a system operator can transfer data stored on a single computer system to all or some of the computer system connected to the first system over a computer network and can do so without expensive electronic server equipment. Moreover, this invention provides the capability of transferring data as files, sectors or cylinders of disk media, thereby permitting a single operator to, through a generally automated procedure, simultaneously install new system software, configuration files and executive files on many computers. This invention provides an important improvement in the operation, maintenance and control of large computer networks, although it applies and works equally well in small network applications. In its best mode of operation this invention is performed on standard digital computer systems through the use of special purpose computer software. Claims We claim: Description BACKGROUND OF THE INVENTION
______________________________________
Command Performed by
Function
______________________________________
Drive Geometry
Master Compare geometry of master image
with slave
RSVP Slave Response to master to indicate
participation in download
Conform Master Acknowledgment that slave has
Download joined the process and that master
knows slave is ready
Sector Data
Master Write data to receive buffer
Sector Data &
Master Write data to receive buffer and flush
Flush data to disk
Skip Track
Master No data in current track - Skip this
track
Resend Request
Slave Slave missed data - Please resend
End of Data
Master Master is finished sending data -
Ready for a slave request
Done Master Image complete - Exit program
Disconnect
Slave Slave response to master "done"
command - Slave disconnecting
Disconnect
Master Master acknowledges slave
Acknowledge disconnect
______________________________________
Current source code for both the IMGBLSTR and the IMGSLAVE programs are included and listed in the Microfiche Appendix included with this patent application. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 depicts a system diagram of a computer network having a source (master) computer and a number of destination (slave) computers connected to each other electronically. FIG. 2 depicts a system diagram of a computer network using the prior approaches utilizing client/server modes of operation for transferring data from one computer to other computers across a network. FIG. 3 depicts the top level state diagram of the master computer ("IMGBLSTR") program/process component of the invention. FIG. 4 depicts the top level state diagram of the slave computer ("IMGSLAVE") program/process component of the invention. FIG. 5 depicts the top level process flow chart of the master computer component of the invention. FIG. 6 depicts the detailed flow diagram of the process menus steps of the master computer component of the invention. FIG. 7 depicts the detailed flow diagram of the upload image step of the master computer component of the invention. FIG. 8 depicts further detail of the read head step of the upload image section of the master computer component of the invention. FIG. 9 depicts further detail of the broadcast head step of the upload image section of the master computer component of the invention. FIG. 10 depicts further detail of the download image step of the master computer component of the invention. FIG. 11 depicts further detail of the fill compress buffer from file step of the down load image step of the master computer component of the invention. FIG. 12 depicts further detail of the get byte from compress buffer step of the down load image step of the master computer component of the invention. FIG. 13 depicts further detail of the write data to head buffer step of the down load image step of the master computer component of the invention. FIG. 14 depicts further detail of the flush head buffer step of the write data to head buffer step of the master computer component of the invention. FIG. 15 depicts a flow chart diagram of the slave component of the invention, which is a detailed flow diagram of the current preferred embodiment of the Slave computer process of the invention. DETAILED DESCRIPTION OF THE INVENTION FIG. 1 shows a computer network having a source (master) computer 101 and a number of destination (slave) computers 102a, 102b, 102c, 102d, 102e, 102n connected to each other electronically as may be used in the peer-to-peer mode of operation of the invention. A typical stand-alone computer may include the following components: a processor; internal memory; a disk storage device; a display device; and an input device. The electrical connection 103 provides a communication channel between the computer systems. The use of this invention does not require that the connection between the computer systems, as designated 103, necessarily be electrical. Other alternative methods of conductivity include, fiber optical, RF, and/or light wave transmission and detection. Often these types of communications channels are referred to as "networks," "local-area-networks" (LANs), and "wide-area-networks" (WANs). Typically, each computer is capable of operating as a stand-alone system. With the addition of the electrical connection 103 the computers are also capable of sharing information (for example data files and e-mail). The applicants' present invention extends the capabilities of individual computer systems connected over a computer network by providing disk imaging from one computer to another computer over a network. Such disk imaging provides a solution to the problem of installing, backing up, maintaining, and upgrading of computer operating system software, computer system utility software, and application programs. This invention, which operates either with or without a computer network server, permits a single computer system to be designated as a "master" and to transfer data from its disk drive or drives to one, some or all of the other computers on the network, designated as "slaves." FIG. 2 depicts a system diagram of a computer network employing the traditional network server hardware 201. This system is well known for providing a method of transferring data to and from the network server and from and to the individual computers on the network. These prior approaches, typically are referred to as "client-server" systems. In a traditional client-server system, data is transferred from a client computer 202 to the network server 201, or is transferred from the server 201 to one or more of the client computers. While applicants' invention supports client-server network systems, the most important achievement of this invention is that it provides the capability of transferring data, files, and disk sectors from any computer on the network to any other computer on the network with or without a network server, thereby dramatically decreasing the difficulty and complexity of computer network management. Where a network server 201 is employed in the network the "master" computer 202a is provided with the capability of transferring data either to the network server 201 and/or to one or all of the "slave" computers 202b, 202c, 202d, 202e, . . . , 202n. This inventions supports networks utilizing client/server modes of operation for compatibility with existing networks. It also supports and provides peer-to-peer operation to permit network management without the considerable expense of network server hardware and simultaneously providing a significant improvement in data transfer efficiency across a computer network. In its best mode of operation, the present invention operates through coordinated use of two computer system communications programs: IMGBLSTR and IMGSLAVE. IMGBLSTR is designed to operate on the "master" computer, while IMGSLAVE is designed to operate on the "slave" computer. In the current preferred embodiment of the invention each program is written in the C programming language and operates on a DOS operating system. However, other equivalent embodiments of the invention may be created in other computer languages, including but not limited to C++, Pascal, and assembly code, and may be designed to operate on other computer operating systems, including but not limited to UNIX, Windows and MacIntosh. The following discussion of the steps of the process of this invention correspond to programs, sub-programs, and routines in IMGBLSTR and IMGSLAVE best mode of the invention. Following this discussion of the essential steps of this process is a complete list of the source code for each program. This source code is provided as part of this disclosure, to provide a fully enabling description of the invention. It is suggested that the reader refer to this source code for additional detailed description as the reader reviews the following figure by figure discussion of the process of the invention. The following description of the top level state diagrams of the "master" computer and the "slave" computer is provided to give the reader an overview of the handshaking and data processing process of the invention. A detailed breakdown of each subprocess, task or program is given in the later detailed descriptions of the process flow charts and the best mode of operation of the invention is provided in the listing of software source code included in this description following the description of the detailed subprocesses, tasks, and/or programs. FIG. 3 depicts the top level state diagram of the master computer ("IMGBLSTR") program/process component of the invention. The master computer process begins by entering the Wait for Slave Connection state 301. In this state 301 the master computer is initialized for communication with slave computers. The process goes from state 301 to state 302 when the master computer broadcasts its disk drive geometry to the slave computers. In this Send Geometry Packet state 302 the master computer sends the master computer drive geometry to slave computers on the send IPX socket. The "currently connected clients" display is initialized. Geometry packets are sent periodically to the slave computers. Geometry packets, in the preferred embodiment of the invention, include the following disk drive information for the master local disk drive: maximum cylinders; maximum heads; maximum sectors, image file partition start cylinder; and image file partition end cylinder. Geometry packets also include information on the IPX network sockets, including: the network address; the node address; and the socket number. Error conditions are monitored and the master computer waits for RSVP packets from the slaves. When the master computer receives an RSVP from the slave the master adds the slave (or client) to its list and send an ACK packet to the slave 303. When the connection between the master and one or more slaves is either completed, terminated by an operator command, or terminated by a timeout, the master process goes to the Send Data Mode state 304. The Send Data Mode state 304, in combination with the Send Head Broadcast state 306 and the Loop to Send All Tracks state 305, performs the upload image, broadcast head, broadcast skip head, broadcast given head, and the send IPX packet tasks. In the event that a Resend Request is received the process goes from the Send Data Mode state 304 to the Queue Req. On Resend List--Perform Throttle Processing 307. During this state, incoming packets are processed, resend requests are processed by servicing the resend queue by sending data packets to slaves, resend requests are recorded, the resend list is processed and tracks are re-transmitted. These tasks are processed in conjunction with state 306. In the event that no more data remains to be sent, the process moves to the Allow Slaves to Catch-Up state 308. This state finishes processing of broadcasts and handles late resend requests. The transition of the process from state 308 to state 309 is accomplished as follows: (1) a timer indicates it is time to send a packet; (2) a packet of data is sent 309; and (3) the process returns automatically to state 308 to wait for the next timer indication. Thus, when the timer has run, the process moves from the Allow Slaves to Catch-Up state 308 to Send "End of Data" Packet state 309, which sends farewell broadcast packets, and back to state 308. When the transfer is complete (when a timeout occurs), the process moves to the Farewell to Slaves state 310, which in combination with the Send "Goodbye" Broadcast state 312 and the Remove Slave From List Send "Goodbye" to Specific Slave state 311, performing the say Goodbye to Slaves and Processing Image Packets tasks. FIG. 4 depicts the top level state diagram of the slave computer ("IMGSLAVE") program/process component of the invention. Beginning with the Hook-Up with Master state 401 the slave computer attempts to form a communications link with the master computer. After a geometry packet is received from the master, the slave process moves to the Send RSVP state 402, where the slave checks the disk drive geometry for compatibility and acknowledges receipt of the geometry information with an RSVP. After the slave process receives a RSVP acknowledge from the master, the process enters the Download Data state 404. Data is received via transitioning between the Download Data state 404 and the Process Data Packet state 407. The Process Data Packet state 407 processes the received data by checking to see if the head is at a valid position, and storing the received data in a buffer. A test is performed to determine if the designated slot is not empty, wherein a "lost data" resend request is sent. The process of the invention provides the capability of sending resend requests through the Send Resend Request state 408. If a "flush" packet is received (a "flush" packet is a special type of data packet, containing an attribute to indicate that when the data is copied to the image file, that the image file should then be flushed to disk), the Flush Complete Track state 405 is entered, where if the buffer is complete, it is written to disk and the "good upto" track is updated, removing the track from the "missing list." If the buffer is incomplete, this state 405 sends a "lost data" resend request and adds the track to the "missing list." When the transfer of data is complete, a "End of Data" packet is received from the master, leading the slave process to the Farewell to Master state 409, where once a "goodbye" is received from the master, the Random Delay--Send "Goodbye" to Master state 410 is entered to send a "goodbye" is send to the master. Thereby, finishing the slave process. FIG. 5 depicts the top level process flow chart of the master computer component of the invention. In its current best mode of operation, the process of this invention begins with the initialization 501 of the master. During this initialization step 501 the user license is verified, usage parameters are checked the MAC address of the computer is acquired and displayed, the local drive geometry is retrieved, all needed buffers are allocated, and the number of hard disk drives installed in the computer is determined. Next, the command line process step is performed 502, where the command line is parsed, setting flags and/or calling the functions to process each command. Next, menus are processed 503. The process menu step calls other functions to get needed information for the execution of the program. A test is made as to whether the process is required to upload data ("image") to a slave or download data ("image") 504. If the process is required to upload data then the upload image step 505 is performed to upload the image, drive data, to the image file. If the process is not required to upload data then the download image step 506 is performed, which accomplishes the downloading of the image (drive data). If a file name was specified by the operator, it is opened and the drive data is read from the file. Lastly, the clean-up step 507 is performed to reset all affected computer systems on the network. FIG. 6 depicts the detailed flow diagram of the process menus steps of the master computer component of the invention. The process menus program process calls other program and/or functions to get the needed information for the process of the invention. Where the requested information has already been entered on the command line, the associated menu is not displayed. First, the program variables are initialized 601. Next, the desired process function is requested 601, allowing the operator to specify whether the operator wants to upload or download. Next, the get partition 603 step is performed to display the local drive partition table and to allow the operator to specify the partition to be imaged. The next step is to get the image file name 604, during which the operator is prompted to provide the name of the image file, including the complete path. Next, the broadcast state is requested 605, that is, whether the operator wants to send the drive data to other slave computers via broadcast packets. FIG. 7 depicts the detailed flow diagram of the upload image step of the master computer component of the invention. The upload image program process performs the function of uploading the image (drive data) from a master computer to a network server computer and/or, if the broadcast feature is enabled, broadcasting the data to the image slave computers on the network. Initially, upload image program is initialized 701. This initialization step is necessary to provide needed data. A test 712 is made to determine whether the invention is operating in the client/server mode. If it is, an image file is opened 702 and an image header is written 703. The drive head buffer is read 704. A test is performed to determine whether the data is to be broadcast 705 to multiple slave computers. If the data is to be broadcast the broadcast is performed 707, if not the process passes through the null state 706 before once again testing to determine if the process is in the client/server mode 713. If it is, the data is compressed and written to a file 708. Data compression is accomplished, currently by a "Run-Length Encoding" scheme well known in the art. Next, System pointers are incremented 709 and if the system has not reached the stop cylinder 710 the process continues to read the head buffer 704, broadcasts the data 707, compresses the data, and adjusts the system pointers. After the stop cylinder is reached 710, the process performs the clean-up step 711 by closing files, resending missed tracks and saying "goodbye" to slave (client) computers. FIG. 8 depicts further detail of the read head step of the upload image section of the master computer component of the invention. This program step reads a head worth of data from a selected local hard disk drive at the current cylinder and current head location. Errors that are encountered, are reported. The process of the invention is aborted after five retries on errors. However, because CRC type errors are recoverable, they will not count towards aborting the process. First, this sub-process is initialized 801, providing access to the necessary variables. Next, the data in the designated current head is read 802. If a read error occurs 803 the read of head data is attempted again, up to five times 805. If errors continue beyond five tries 805 an error message is displayed 806 and the program is exited 807. Otherwise, if the data is read without error then the process passes through the null state 804 and the program returns to the upload image process step 505 as shown in figure seven. FIG. 9 depicts further detail of the broadcast head step of the upload image section of the master computer component of the invention. The function of the broadcast head program--subprocess is to broadcast the drive head data (headbuffer) to the image slaves using the send IPX socket. Data is sent 512 bytes at a time, equal to a single sector of disk data. On the last sector sent, the command to flush the data, that is, to write it to the hard drive, is sent. The first step in the broadcast head process is to initialize the data for broadcast 901. The process of sending packets of data 905 is performed one sector at a time. After it is determined that an additional sector of data is in the head 902, the sector data is packaged 905 or prepared for broadcast, a delay is provided 903 to insure correct interpacket data broadcast, then the packet is sent 905. The broadcast head process is finished when no further sectors remain for transfer. FIG. 10 depicts further detail of the download image step of the master computer component of the invention. This program sub-process of the invention performs the function of downloading the image (disk drive data). If a file name is specified by the operator, it is opened and the disk drive data is read from the file. If the broadcast feature is enabled, the data is broadcast to the image slaves on the network. The first step is to initialize 1001 the needed data and to set the head buffer pointer to 0. Next, an image file is opened 1002 if needed and if the image file name is valid. The image header data is read 1003 and a check is made for the proper drive geometry. If the image geometry is equal to the drive geometry 1004 then the data is decompressed 1006 using well known "Run-Length" decompressions schemes. Once the data is decompressed it is written 1007 to the head buffers. After data is written to the head buffer, this part of the process is complete. If the image geometry does not match the drive geometry, of step 1004, an error message 1005 is displayed. FIG. 11 depicts further detail of the fill compress buffer from file step of the down load image step of the master computer component of the invention. This function fills the compress buffer by reading a head size (maximum sectors times 512) of data from the image file. If an error occurs during the read, an error message is displayed. First, the data is initialized 1101 to provide data for the process. Next, the compress buffer file size is read 1102 and tested 1103 for error. If an error is encountered, it is displayed 1105, otherwise a null state 1104 is used and the process returns to the download image process of the invention of FIG. 10. FIG. 12 depicts further detail of the get byte from compress buffer step of the down load image step of the master computer component of the invention. This process subprogram functions to retrieve a byte from the compress buffer. When the compress buffer is empty, it is refilled by reading data from the image file. First, the data is initialized 1201 for use in the process. Next the compress pointer is tested 1202 to determine if it is greater than the size of the compress buffer. If it is, the compress buffer is filled 1203 from the image file. The compress pointer is then reset 1204. The byte pointed to by the pointer/counter is returned 1206 and the process returns to the download image process of FIG. 10. In the event that the compress pointer does not exceed the size of a buffer, test step 1202, the process goes through a null state 1205 and returns 1206 the byte pointed to by the compress pointer and returns to the download image process. FIG. 13 depicts further detail of the write data to had buffer step of the down load image step of the master computer component of the invention. This step functions to take the byte of data passed to it and write it to the head buffer the number of times indicated by the compress counter. This is part of the process of decompressing the image file. When the head buffer is full, it flushes the data to the hard disk drive. First, the data required is initialized 1301. Next, the data is written into the head buffer 1302. The head buffer pointer is incremented 1303. A test 1304 is made to determine whether the head buffer is full. If it is, the head buffer is flushed to disk 1305 and the pointers/counters are adjusted 1306. A test 1308 is then made to determine whether the compress count is complete. If it is, the process returns to the download image of FIG. 10. If it is not, then the process returns to step 1302 to move a byte into the head buffer. If the head buffer, of step 1304, is not full, a null state 1307 is passed through before the compress count test of step 1308. FIG. 14 depicts further detail of the flush head buffer step of the write data to head buffer step of the master computer component of the invention. This step functions to flush the head buffer data to disk. The entire head buffer is written in one command. All needed pointers, counters, etc., are updated along with screen information. Also, if the broadcast feature is enabled, the head data is broadcast to the image slaves. First the data is initialized 1401 for use in this process. Next, a test 1402 is made to determine if the process is in the broadcast mode. If it is, the head data is broadcast 1403 to all image slave computers. If not, the process goes through a null state 1404. The head buffer is then written 1405 to disk and the current head and cylinder data is adjusted 1406 prior to returning to the write data to head buffer process of FIG. 13. FIG. 15 depicts a flow chart diagram of the slave component of the invention. is a detailed flow diagram of the current preferred embodiment of the Slave computer process of the invention. The IMGSLAVE is the slave component or process of the invention that provides the parallel disk image process. The slave uses an IPX socket to listen for data from the image master. A special header in the data from the master determines the function the slave will perform. The IMGSLAVE program cannot create or restore an image without the IMGBLSTR program. Its function is to listen to the network for data from the IMGBLSTR, to retrieve data from the network and to write it to the slave's local drive. The slave process begins by initializing 1501 data for processing. When a geometry packet is received 1502 from the master the slave responds with an RSVP 1503. Next, the slave listens for an RSVP acknowledge 1504 from the master. After the receipt of the RSVP acknowledge, a test 1505 is made to determine whether the register for download is valid, if not, a message indicating that the transfer is unusable is sent 1506. Alternatively, if the register for download is valid, then the data is downloaded 1507 to the slave. Error checking 1508 is performed and errors are displayed 1509 if detected. If no errors are detected, a download complete message is displayed 1510. The following is a listing of the computer source code which is the current best mode preferred embodiment of the invention. The reader can, by consulting this source code, learn all that is necessary to produce and use the invention. The previous description, including the listed source code, describes the current preferred best mode embodiment of the invention as it is performed on personal computers connected through a computer network with or without a computer server. The software programs which are used to practice this invention typically reside in the memory and/or hard disk storage medium of the networked computers. While the current best mode of the invention is used on personal computers, it is not necessary that it be limited in this way. Any computational device which has a long term storage medium, for example: a disk drive, a tape drive, a CD or optical storage medium; and can be networked to other computational devices. The size, configuration or purpose of the computational device does not limit the use of this invention to image long term stored data from one computational device to another in either a peer-to-peer mode or a client/server mode of operation. Furthermore, while this invention is performed, in its current best mode, by software written in the C programming language, alternative computer languages can equivalently used. The software source code provided as part of the disclosure of this patent application, shows in detail how the functional steps of the invention are performed. Of course, it is contemplated that the inventive concept of this invention may be implemented through other techniques and in other embodiments and in other computer languages. The computer source code is provided to describe the best mode of operation of the invention, such a best mode may evolve and change over time, after the filing of this application without altering the fundamental inventive concept of the method, which is the imaging of computer data from one computer to one or more others over a network using a peer-to-peer method and still remain compatible with a client/server mode of operation, without requiring special purpose network server hardware.
|
Same subclass Same class Consider this |
||||||||||
