NuoDB Swifts Release 2.1

NuoDB Swifts Release 2.1

InDetail Paper by Philip Howard, Research Director/Information Management, Bloor Research.

–Date October 2014

Key findings

In the opinion of Bloor Research, the following represent the key facts of which prospective users should be aware:

• NuoDB has cracked the problem of how to distribute a database across multiple physical locations (which may be geographically separated and/ or in the cloud) in order to support transaction processing in a familiar SQL-based environment while providing performance, scalability and relatively low cost.

• Despite the fact that NuoDB presents a SQL face to the world it is not actually a relational database under the covers. However, you won’t see that and users don’t need to know about it. Under the covers the product uses a distributed object model and NuoDB is actually a multi-model database with the company planning to introduce additional interfaces during the course of 2015.

• Data may be stored in flat file systems or in key-value stores such as HDFS or Amazon S3. Thus storage is relatively low cost. This also implies, correctly, that the product is available both as in-cloud and as an on-premise solution, or in hybrid environments.

• A typical environment consists of multiple NuoDB domains within which each storage manager has its own copy of the (same) database. Processing is logically distinct from storage and is performed by transaction engines based on a memory-centric architecture that performs at in-memory speeds if all the data is in memory. You can have as many or as few transaction engines and storage managers within a single domain as you wish, and there is no necessary relationship between the numbers of each now offers HTAP (hybrid transaction/ analytical processing). This is enabled by allocating transaction engines and storage managers within the cluster to do analytic tasks that do not interfere with operational processes.

• NuoDB is an append-only database that makes extensive use of multi- version concurrency control (MVCC). Corollaries to this approach mean that you can make point-in-time enquiries (for compliance purposes, say) and that the product supports

• The database is ACID compliant and provides support for concurrency, locking and commit processes, albeit in a somewhat different way from traditional approaches.

• The database is continuously available (that is, regardless of planned or unplanned downtime) and it is easy (and mostly automatic) to scale the environment up or down (there are facilities to quiesce database instances) on demand.

DOWNLOAD FULL REPORT (.pdf): Bloor-In-Detail-NuoDB

This document is copyright © 2014 Bloor.

You may also like...