Comparing 1010data and Hadoop (and RDBMS)
It is tricky to directly compare 1010data with Hadoop since it is an apples and oranges comparison. 1010data is a complete solution while Hadoop is merely one relatively small component of a solution. Hadoop is a data management engine, while a complete solution must also include database development and management, user interfaces, analytics, and so on. To use an analogy, if 1010data is FedEx, Hadoop is a truck motor. The two are unlikely to be viewed as competing options.
A more meaningful line of inquiry is to compare 1010data with a complete Hadoop-based solution, i.e. a system that uses Hadoop as its database engine but that also includes all the other components that comprise a complete solution. Let's then do this, and while we're at it, throw in relational database managers (RDBMS) as well. Green is good, red bad, orange somewhere in the middle.
|Effort (Implementation and Maintenance)|
|Time to Value|
|Simplified Data Integration|
|Data Mash Ups|
|Real-Time Updates (OLTP)||5|
- Covers software RDBMS, database appliances, in-memory databases, and columnar databases. These alternative technologies have different cost and performance characteristics, but the differences are small relative to this analysis.
- In theory Hadoop is open source and therefore free, but most companies will opt to purchase a supported version. More importantly, there is a costly human effort involved and there are significant charges for other system components.
- In the case of all three options, and especially Hadoop, performance is highly dependent on the number of nodes. For this chart we assume that the number of nodes in each of the three solutions are within, say, an order of magnitude of each other.
- With enough time and effort, Hadoop can be made to support advanced analytics, however the cost is usually prohibitive, ease-of-use is extremely limited, and performance will always lag other options.
- 1010data is an analytical platform meant for analysis and operational reporting. It is not designed for transaction processing.