Rule of thumb for sizing
How much main memory is required?
System sizing is based, on the one hand, on analyzing user behavior and, on the other hand, on the properties of your existing systems. In the following, an example is used to explain the sizing procedure.
In a 10-TB data warehouse, a large proportion of the data are irrelevant, in terms of content, to the result (e.g. indices, materialized views or speculative aggregates). These support structures are superfluous for the EXASolution system and are not taken into account when estimating the size of the database.
Our many years of experience show that existing databases include historical data that occupy a large amount of space and that are never or very rarely accessed.
However, active data are data that are accessed more frequently in response to queries. EXASolution achieves an optimum performance when all active data are compressed and stored in the system’s working memory.
As a result of EXASolution’s efficient data compression, the system’s main memory requirement is again reduced. Thus, the estimated 0.5 TB of active data in a cluster of 20 nodes, each with 32 GB RAM, can be processed to optimum effect.



