Data Corruption

About: Data Corruption is a research topic. Over the lifetime, 435 publications have been published within this topic receiving 6784 citations.

Memory including portion storing a copy of primary operating system and method of operating computer including the memory

Abstract: A method of installing an operating system into a computer entity comprising at least one data storage device, a primary operating system and an emergency operating system, the method characterized by comprising the steps of: creating a copy of the primary operating system on an operating system back-up area of the data storage device of the computer entity; operating the computer entity using the emergency operating system; storing data of the computer entity on a user settings archive of the data storage device; replacing the primary operating system with the copy of the primary operating system; automatically checking for corrupted data on the user settings archive; restoring settings data of the computer entity from the user settings archive. In the event of a failure involving data corruption, application data may be deleted and recreated in a known good default state.

Pangolin: A Fault-Tolerant Persistent Memory Programming Library

Abstract: Non-volatile main memory (NVMM) allows programmers to build complex, persistent, pointer-based data structures that can offer substantial performance gains over conventional approaches to managing persistent state. This programming model removes the file system from the critical path which improves performance, but it also places these data structures out of reach of file system-based fault tolerance mechanisms (e.g., block-based checksums or erasure coding). Without fault-tolerance, using NVMM to hold critical data will be much less attractive. This paper presents Pangolin, a fault-tolerant persistent object library designed for NVMM. Pangolin uses a combination of checksums, parity, and micro-buffering to protect an application's objects from both media errors and corruption due to software bugs. It provides these protections for objects of any size and supports automatic, online detection of data corruption and recovery. The required storage overhead is small (1% for gigabyte-sized pools of NVMM). Pangolin provides stronger protection, requires orders of magnitude less storage overhead, and achieves comparable performance relative to the current state-of-the-art fault-tolerant persistent object library.

Multipoint database synchronization protocol to avoid data corruption

Abstract: A database multipoint synchronization which allows multiple clients to simultaneously access and edit a database while avoiding inadvertent data corruption and ensuring the integrity of data within the database. A database manager, which may be configured as part of a database software application, keeps track of modifications saved to a database file and increments a modification change counter accordingly. When a client user accesses a database record, the database manager detects the modification change counter value. Then if that client seeks to save modifications to the database record, the database manager detects the current modification change counter value to discern whether other clients have saved modifications to the record following the access by the client presently seeking to save modifications. If the modification change counter has incremented, the client is denied authorization to save the modifications and offered a choice of alternative operations. In this way, the data within the database record is not corrupted due to inadvertenly overwritting by another client's record.

Detecting silent data corruption through data dynamic monitoring for scientific applications

Abstract: Parallel programming has become one of the best ways to express scientific models that simulate a wide range of natural phenomena. These complex parallel codes are deployed and executed on large-scale parallel computers, making them important tools for scientific discovery. As supercomputers get faster and larger, the increasing number of components is leading to higher failure rates. In particular, the miniaturization of electronic components is expected to lead to a dramatic rise in soft errors and data corruption. Moreover, soft errors can corrupt data silently and generate large inaccuracies or wrong results at the end of the computation. In this paper we propose a novel technique to detect silent data corruption based on data monitoring. Using this technique, an application can learn the normal dynamics of its datasets, allowing it to quickly spot anomalies. We evaluate our technique with synthetic benchmarks and we show that our technique can detect up to 50% of injected errors while incurring only negligible overhead.

Using Checksums to Detect Data Corruption

Abstract: In this paper, we consider the problem of malicious and intended corruption of data in a database, acting outside of the scope of the database management system. Although detecting an attacker who changes a set of database values at the disk level is a simple task (achievable by attaching signatures to each block of data), a more sophisticated attacker may corrupt the data by replacing the current data with copies of old block images, compromising the integrity of the data. To prevent successful completion of this attack, we provide a defense mechanism that enormously increases the intruders workload, yet maintains a low system cost during an authorized update. Our algorithm calculates and maintains two levels of signatures (checksum values) on blocks of data. The signatures are grouped in a manner that forces an extended series of block copying for any unauthorized update. Using the available information on block sizes, block reference patterns and amount of concurrently active transactions in the database, we calculate the length of this chain of copying, proving that the intruder has to perform a lot of work in order to go undetected. Therefore, our technique makes this type of attack very unlikely. Previous work has not addressed protection methods against this knowledgeable and equipped intruder who is operating outside the database management system.