AAEON, an industry leader in edge computing solutions, collaborated with KingTiger, a leader in memory testing solutions, to help with testing and developing their intelligent Memory Surveillance (iMS) memory error prevention software. Developed in conjunction with AMI, a leading provider of firmware, manageability and security solutions for the global computing industry, this innovative software, which is embedded in Aptio V UEFI Firmware from AMI, allows systems to avoid memory errors without ECC support, ensuring secure, reliable, and accurate data reporting in edge computing and big data systems.
Data integrity is vital for a wide range of big data and data-based applications. From medical records to finance, memory errors can wreak havoc on accuracy, reliability and even data safety and security. For larger and more powerful computing systems, this is usually solved with the use of error-correcting code (ECC) memory. However, ECC memory is often limited based on whether or not the processor chipsets support this hardware solution. This often requires developers and users to upgrade their hardware platforms to higher performance, and thus more expensive, platforms where the extra power and cost are otherwise unjustified, such as in edge computing applications.
KingTiger, specialists in memory testing solutions, recently developed iMS, a software-based approach to error correction memory. The iMS, or intelligent Memory Surveillance, software is deployable on platforms which otherwise don’t support ECC memory, allowing developers and users to use platforms with more appropriate processing capabilities, saving costs while maintaining data integrity. The software can detect erroneous bits and mask them before they cause errors within the data and avoid system crashes. iMS provides an advantage over ECC memory in that it can detect and mask an unlimited number of bits, whereas ECC memory is typically limited to one bit per address. It also, in effect, extends the lifecycle of memory/RAM modules by being able to handle a greater number of error bits.