WhiteCanyon Software selectively pursues intellectual property protection on methods and developments in the data erasure industry. This patent protects the process of wiping data on a computer that is reported lost or stolen. This patent reaffirms WhiteCanyon Software as the world leader in data erasure technology. The patent material can be found here:
The ‘selective storage device wiping system and method’ patent provides protection on the methodology for the activation of the wipe process from a remote server or other site when a computer is reported lost or stolen. The wipe technique selectively wipes all data files and free space and then wipes the entire storage device. This causes any personal data files, photos, and videos to be wiped first. Once the wiping process starts, it will continue until complete. The process starts or continues whenever the computer is booted. Only when all personal and sensitive data is wiped, will the system then change the encryption key (if applicable) and begin wiping the entire storage device.
Description of Prefered Embodiments
The present invention relates to a system and method for preventing loss of personal information and other critical files from a lost or stolen target computer that includes assigning a unique ID to the target computer and storing this ID on a remote control server computer in a database accessible by the remote server. The database does not need to be co-located with the remote server, but should be accessible by it at least over a network. The control server is typically located at a control center that manages many different user computers. The target computer communicates with the control server over the network, or by any other method.
A typical embodiment of the invention can be seen by turning to FIG. 1. The target computer's processor communicates via a communications interface with a remote server over a network. The remote server has access to a database that is either stored on a physically connected device or that resides somewhere else in the network. The target processor also has a direct connection with a computer storage device controller (which is usually part of an I/O chip mounted on the motherboard). This controller is connected to the storage device hardware controller that is part of the storage device hardware5. Data and commands travel over buses from the processor to the storage device.
As long as the target computer is not reported lost or stolen, a status is assigned to its ID as “safe” in the database (or some similar status indicator). If the computer is lost or stolen, its status is changed to “unsafe”. This can be accomplished by the control center receiving a report from the owner of the computer either by telephone, online or by any other method, or in some cases, the computer may be tracking its own location and know that it has been stolen. The person or computer making the report can be identified as having authority by password, codeword or by any other verification method.
A small software, firmware or hardware module can be placed on the target computer that runs every time the target computer is booted. This module can also run periodically at a time interval set by the owner or by the control center. This module transmits the target computer ID to the control server over the network and requests the status of the computer. The control server checks its database and returns a status of safe or unsafe.
As long as the status comes back safe, the module ends (or returns to a background state) and turns control over to normal operating system or Basic Input/output System (BIOS) type software. To run periodically, before turning over control to the normal system, the module can set a timer for a high priority interrupt or the like to occur the next time it is to run. Upon this timer interrupt, either the module is re-initialized and started or if simply dormant, it is awakened.
If the status comes back unsafe, the module can become autonomous and begin wiping the target computer's storage device in a very selective manner. For extra security, the module can double check with the control server over the network to make sure the status has not been received in error. For high reliability, the control server can send an encrypted word or command that can only be decrypted by that target computer (such as by a special key or a public/private key system). For example, the server could store the target computer's public key in its database along with the ID and status. A wipe command could then be encrypted using the target computer's public key and transmitted. Only the correct target computer could decrypt the wipe command using its private key. While this extra security is optional, it is preferred since it prevents any accidental wiping that was unintended both on the particular target computer or on another computer that received the command in error.
An status message indicating unsafe can cause wiping of the entire storage space as previously described, or optionally can contain a list of files or folders to be removed without a full wipe. The following are the possibilities for an unsafe status:
- Perform a full wipe, overwriting all user data files and free space first.
- Overwrite all user data files and free space without performing a full wipe.
- Overwrite only the following folders and/or files.
Once the wipe module has determined to wipe, it has no further need to communicate with the control server. However, it can optionally send a message to the control server notifying it that wiping has started. If the target computer is powered down, the module will start where it left off the next time the target computer is booted.
While it is possible in some embodiments of the invention to build in the ability for the control server to stop the wiping operation, this leaves a security trapdoor that can defeat the system. It is therefore preferred that once wiping is authorized, there is no way to stop it, and there is no further communication with the control server.
FIG. 2 shows a message flow diagram between the target computer and the remote server over a network. Upon power up or any other boot condition, a message26 containing the target computer's ID is transmitted to the remote server. The remote server checks its database to find the status of the target computer. The status safe or unsafe is returned from the database to the remote server. The remote server then sends a response message over the network back to the target computer. This response message contains the status or an indication of the status. If the status is safe, the target computer turns control over to the next normal step in its startup sequence. If the status is unsafe, the target computer20 begins, or continues, a wipe operation on the storage device.
The wipe operation is carried out selectively by first either choosing files from an ordered, or unordered, file list placed on the target computer by either the owner/user or by the control server, or by following a predetermined file order. Generally, it is desirable to first wipe all user files. These are any type of file that was created or downloaded on the target computer by the user or put on the computer at initial software loading time. These include document files, text files, photos, videos, spreadsheets, and any other type of file that can be created by any application (on a Microsoft system, this generally includes any file that does not have a .exe, .dll or other system or network file suffix). In other words, all of the files that might contain any information considered personal by the owner are overwritten first.
Next, free space should be overwritten since free space usually contains numerous fragments of user files including temporary buffers, parts of deleted files and the like. If free space were not wiped, a significant amount of user data would remain. Next, the module can optionally wipe all applications (which are executable files) and, if desired, part or all of the operating system (including the registry on machines running Microsoft Windows operating systems). Whether, applications and/or the operating system are wiped can be determined by the user's file list if one is used.
Finally, the module optionally changes the storage device's encryption key (if there is one) and begins to perform a normal wipe of all remaining storage device space (or of all storage device space from the beginning).
As previously stated, prior art wipe systems simply began wiping addresses starting from the lowest to the highest address. These systems had no need to know anything about the file system in use. However, for the wipe module to selectively wipe on a file-by-file bases, it must be able to identify those addresses that belong to a particular file. In order to do this, it must first identify what type of a file and directory system is being used, be able to read the directory, and from the directory determine what addresses belong to a particular file. There are two ways it can do this: 1) the wipe module can make system calls to the operating system (if the operating system is running) to retrieve addresses, or 2) it can operate independently of the operating system. The latter method is preferred; however, it makes the module somewhat more complex. The main reason for not depending on the operating system is that the module ideally runs at boot time before the operating system is brought up.
Also, the module typically must contact the remote control server over the network at least once when it starts. This can be done using the capabilities of the operating system to perform network communication (using the standard communications stack); however, direct communication can also be performed over the network by the module. To do this, the module must first determine what type of communication hardware the target computer has (wireless, plugged-in Ethernet™ or the like). It must then set up communications onto the network (for example by using a wireless interface) and send the correct sequence of commands over the network to the control server. While this is more difficult to implement, it is also more secure. As stated, once, the module has been given a wipe command by the control server, it generally cannot be reset. This assures that it is very difficult to disable it once it has started.
For the wipe module to independently access the communication interface without the help of the operating system, it is necessary to be able to identify the network interface and to operate it. This can be made much easier by supplying the wipe module with a link to the a communications driver program when it is loaded onto the particular target machine. If that is done, the wipe module needs only to follow standard interface rules for all drivers, and does not have to identify the actual communications hardware or be concerned with what type of interface it is.
Of course, a knowledgeable thief could attempt to thwart the module by taking measures to prevent network communication until the desired data had been removed. To avoid this, the module can determine that it is unable to contact the system server to obtain the status of the computer. In this particular case, the module can go into an undecided state, and while not actually starting a wipe operation, it can prevent normal operation of the computer until it can contact the control server.
The case where the wipe module is an application and is loaded into random access memory (RAM) and runs under the operating system is shown in FIG. 3. The operating system (or parts of it) are loaded into Random Access Memory (RAM) memory (typically using virtual addressing). The wipe module in this mode can run as a privileged application that is automatically started after the operating system boots (and can also run periodically using a timer). The wipe module accesses the network though a communications controller that interfaces with a network physical device. The network access can be managed by commands to the operating system using the standard communications stack and running as a high priority application. Access to the storage device and the file directory stored on the storage device is through normal operating system commands. These commands can be high level allowing the operating system to manage the directory and file interface, or low level using device storage addresses. Actual wipe writes to addresses can be low level; however, the wipe module, running as an application, must have a high enough privilege level to allow it to directly write to the storage device addresses at the address level.
In the case that the module does not rely on the operating system (which is a more secure mode) where it has determined that it must wipe, the module takes control of the computer after boot. On its first execution, it can determine what type of file system is present (FAT, NTFS, etc.), read the main directory, and begin to execute without ever turning control over to the operating system or allowing the operating system to boot. It can parse the main directory (and hence sub-directories), searching for user files and wiping their associated addresses file by file as previously described.
FIG. 4 shows a block diagram of an embodiment where the wipe module resides in read only memory (ROM), or some other form of permanent or semi-permanent memory along with boot firmware such as a BIOS. The wipe module can use the same part of RAM memory for buffers and the like as the boot firmware. In this case, BIOS commands can be used to interface with the communications controller and the physical network interface. Again, a link to a driver module can be supplied to the wipe module when it is loaded on a specific target machine. The wipe module can either directly interface with the storage device controller or use BIOS commands to interface. Again, the wipe module must be able to determine the type of file system used on the storage device and read and use the directory. The wipe module can then selectively wipe user files, followed by free space, followed by a general wipe as explained above.
The most secure embodiment of the invention does not use the computer's processor at all. Rather, the wipe module is stored and runs on the storage device's controller hardware/firmware (or is implemented as an independent hardware device). This embodiment generally does not communicate with a control server. Rather, it simply determines if the storage device has been moved to a different computer. If it has, the module begins a selective wipe, and will not allow normal storage device operation to take place. This particular embodiment of the invention is effective when a storage device is removed from one computer and placed in another (perhaps to avoid a computer software/firmware module that would perform a wipe). Typically, this embodiment is used for high-security storage devices. When the storage device is first placed in a computer, the wipe module is in a passive state. In this state, it determines what computer it is residing in. It does this by reading a code or number over its communication bus. The storage device then records this computer ID, and continues to stay passive as long as the storage device resides in that computer. Every time the computer is powered up, it again determines what computer it is in. If the computer is different, or if it cannot determine a computer, it begins the selective wipe operation.
There are several possibilities for this embodiment:
The storage device determines what computer it is in by reading an ID from the computer.
The computer sends the storage device an ID code that must be received before normal storage operation can take place. This code can then be stored in the storage device controller (optionally encrypted). If it finds itself in a different computer, wiping can begin.
Optionally, in some embodiments, the storage device can pretend to function normally (to fool a thief), but in reality be wiping.
FIG. 5 shows a block diagram of an embodiment that runs on a storage device hardware system. A hardware storage device controller interfaces with a ROM memory and a RAM memory. This is generally part of the storage device or device hardware that is supplied by the device manufacturer. Therefore, this embodiment generally requires cooperation of the storage device manufacturer. Storage device control firmware along with the wipe module is stored in the ROM, and is execut by the controller. The controller communicas with the computer the storage device is installed in over a standard interface. The storage device controller can determine a computer ID58 over that as has been described. The device controller can thus determine if the storage device has been moved to a different computer. If so, a wipe operation can begin. Again, the wipe module must determine the type of file system and read the directory. It can then selectively wipe addresses belonging to data files that in the order previously discussed. In this embodiment, once the wipe begins, it typically cannot be stopped.
The present invention provides a very rapid way of getting rid of important sensitive personal data and files on a storage device that is marked as unsafe either by being reported stolen, or in some embodiments by having the storage device moved to a different computer. The present invention is much faster than prior art techniques that attempt to wipe the entire storage device without regard to the file structure or content.
Several descriptions and illustrations have been given to aid in understanding the present invention. One with skill in the art will realize that numerous changes and variations may be made without departing from the spirit of the invention. Each of these changes and variations is within the scope of the present invention.
The patent for a selective storage device wiping system is just one of the erasure patents that WhiteCanyon Software maintains. The patents further reassert WhiteCanyon’s role as the leader in the data sanitization industry. For more information on WipeDrive Enterprise and its implementation in your organization, call 801.224.8900.