Sunday, November 13, 2011

Dynamic Random Access Memory as a possible replacement for Hard Disks

Dynamic random-access memory (DRAM) stores each bit of data in a separate capacitor within an integrated circuit. It is structurally simple (only one transistor and a capacitor are required per bit) allowing DRAM to reach very high densities. The transistors and capacitors used are extremely small; billions can fit on a single memory chip. However, the memory is volatile and data is lost when power is removed.

There is a a research group at Stanford University who say the goal should be to replace hard disks with DRAM. RAMcloud is a general-purpose storage system where all data lives in DRAM at all times and large-scale systems are created by aggregating the main memories of thousands of commodity servers. RAMCloud provides durable and available DRAM-based storage for the same cost as volatile caches, and it offers performance 10-100x faster than existing storage systems. By combining low latency and large scale, RAMCloud will enable a new class of applications that manipulate large datasets more intensively than has ever been possible.

With increaed throughput (1m ops/sec/server) and low latency access the implication is that we will see a new class of applications that operate large scale using very large datasets. Possible applications are: 
  • crowd level collaboration;
  • large scale graph algorithms;
  • real-time information intensive applications
  • applications using open linked data 
There are a number of issues that are being addressed by the research group including durability and availability, the right data model, concurrency and consistency and keeping up to date on the progress will be extremely prudent.

A useful introductory lecture by John Ousterhout is Professor (Research) of Computer Science at Stanford University can be watched here.