A survey of memory error correcting techniques for improved reliability

Computer systems failure due to hard or soft memory errors is very common. Hard errors are caused due to any permanent fault in the memory chips whereas soft errors in memory chips, generally transients or intermittent in nature, are caused due to alpha particles or cosmic rays. Non-critical systems may not require serious attention for such failures where simple, cost-effective, little-overhead techniques may be considered enough. However, semi/fully critical systems do require a careful treatment, keeping aside all other factors, but the reliability and serviceability during the intended period of time. A number of monolithic and hybrid techniques have been developed over the years. This paper aims to expose the concerns related to increasing memory and logic errors as an off-shoot due to the advancement in technologies. A survey is presented regarding the techniques being used to deal with such errors.