The case for RAMCloud
Communications of the ACM
With scalable high-performance storage entirely in DRAM, RAMCloud will enable a new breed of data-intensive applications. by John ousterhout, Parag agrawal, david erickson, christos kozyrakis, Jacob leverich, david mazières, subhasish mitra, aravind narayanan, diego ongaro, guru Parulkar, mendel rosenblum, stePhen m. rumble, eric stratmann, and ryan stutsman For the past four decades magnetic disks have been the primary storage location for online information in computer systems. Over that
... d, disk technology has undergone dramatic improvements while being harnessed by higher-level storage systems (such as file systems and relational databases). However, disk performance has not improved as quickly as disk capacity, and developers find it increasingly difficult to scale disk-based systems to meet the needs of large-scale Web applications. Many computer scientists have proposed new approaches to disk-based storage as a solution, and others have suggested replacing disks with flash memory devices. In contrast, we say the solution is to shift the primary locus of online data from disk to DRAM, with disk relegated to a backup/archival role. A new class of storage called RAMCloud will provide the storage substrate for many future applications. RAMCloud stores all of its information in the main key insights the web has driven development of new large-scale applications that have effectively scaled compute power and storage capacity but have not scaled storage access rates to match. dram-based storage can be made as durable and available as diskbased storage without giving up its performance advantages. dram-based storage is not just faster than disk or flash; for i/o-intensive workloads it is also cheaper and more energy efficient. are all affiliated with the Department of computer science of stanford university, stanford, ca.