Stop Thinking, Just Do!

Sungsoo Kim's Blog

Main Memory DBMS

tagsTags

28 January 2016


Article Source

  • Title: Main Memory DBMS
  • Authors: PETER BONCZ, CWI, Amsterdam, The Netherlands

Main Memory DBMS

Synonyms

In-memory DBMS; MMDBMS

Definition

A main memory database system is a DBMS that primarily relies on main memory for computer data storage. In contrast, conventional database management systems typically employ hard disk based persistent storage.

Key Points

The main advantage of MMDBMS over normal DBMS technology is superior performance, as I/O cost is no more a performance cost factor. With I/O as main optimization focus eliminated, the architecture of main memory database systems typically aims at optimizing CPU cost and CPU cache usage, leading to different data layout strategies (avoiding complex tuple representations) as well as indexing structures (e.g., B-trees with lower-fan-outs with nodes of one or a few CPU cache lines).

While built on top of volatile storage, most MMDB products offer ACID properties, via the following mechanisms: (i) Transaction Logging, which records changes to the database in a journal file and facilitates automatic recovery of an in-memory database, (ii) Non-volatile RAM, usually in the form of static RAM backed up with battery power (battery RAM), or an electrically erasable programmable ROM (EEPROM). With this storage, the MMDB system can recover the data store from its last consistent state upon reboot, (iii) High availability implementations that rely on database replication, with automatic failover to an identical standby database in the event of primary database failure.

Main-memory database systems were originally popular in real-time systems (used in e.g., telecommunications) for their fast and more predictable performance, and this continues to be the case. However, with increasing RAM sizes allowing more problems to be addressed using a MMDBMS, this technology is proliferating into many other areas, such as on-line transaction systems, and recently in decision support. Main memory database systems are also deployed as drop-in systems that intercept read-only queries on cached data from an existing disk-based DBMS, thus reducing its workload and providing fast answers to a large percentage of the workload.

Examples

Examples of main-memory database systems are MonetDB, SolidDB, TimesTen and DataBlitz. MySQL offers a main-memory backend based on Heap tables. The MySQL Cluster product is a parallel main memory system that offers ACID properties through high availability.

  1. Bohannon P., Lieuwen D.F., Rastogi R., Silberschatz A., Seshadri S., and Sudarshan S. The architecture of the dali main-memory storage manager. Multimedia Tools Appl., 4(2):115–151, 1997.
  2. Boncz P.A. and Kersten M.L. MIL primitives for querying a fragmented world. VLDB J., 8(2):101–119, 1999.
  3. DeWitt D.J., Katz R.H., Olken F., Shapiro L.D., Stonebraker M., and Wood D.A. Implementation techniques for main memory database systems. In Proc. ACM SIGMOD Int. Conf. on Man- agement of Data, 1984, pp. 1–8.
  4. Hvasshovd S-O., Torbjornsen O., Bratsberg S.E., and Holager P. The ClustRa telecom database: high availability, high throughput, and real-time response. In Proc. 21th Int. Conf. on Very Large Data Bases, 1995, pp. 469–477.

comments powered by Disqus