BACKGROUND TO THE FOLLOWING NEW ENGINE FEATURES
As Alterian Campaign Manager continues to be developed to meet the diversity of our Client databases, Campaign Manager 6.0.8 contains the following new features which are for specific purposes, data volumes and usage patterns and in some cases are initially implemented with limited scope.
As such these features have been deemed as not suitable for enabling by default, so are initially disabled and implementation should only be carried out after discussion via Alterian Support & Customer Services to assess their suitability.
• B-TREE COLUMN INDEXING
An additional index method available in Engine, although currently only implemented on High Cardinality VARUNICODE column types. This index storage has a high level of abstraction, meaning it is tuned for small, rapid changes to data i.e. accessing columns in Campaign History where an Intra-Day update strategy is implemented. The new index has been fully integrated into existing query, engineering functions, making it fully functional but, based on data type, this feature has limited in scope at this time.
• MIRRORED LINKING
In current SEV processing a “lazy” or “On demand” approach has been implemented for the creation of links after a storage swap of mirrored table processing occurs. Link key modifications are tracked but not processed, so links are re-created on demand via the user performing, for example, a Cross table Query or Campaign History crosstab. This feature performs the link creation as part of the “behind the scenes” process immediately after the storage update, but before the storage swap and publication of the new data in the Campaign History tables. This means the link processing would delay the publication of data, but remove any lag in user experience while the new link is built on demand. Usage is dependent on factors such as frequency of access to the tables and data volumes between storage swaps, and again only advised after discussion via Alterian Support & Customer Services to assess their suitability.
NEW CERBERUS.INI SETTING
The following settings have been created to exist alongside the existing AggregateRealWidth, Public setting to complete the functionality across the necessary data types. These settings allow control over the width of expressions yielding the defined data type.
Please note the enhancement defined below in regards to the new default data type for Count Aggregates to BIGINT, as these settings can be used to define the Aggregate width should this need to be consistent with previous versions.
- ExpressionNullWidth - Default = 0
- AggregateBigIntWidth - Default = 20
- ExpressionBigIntWidth - Default = 20
- ExpressionDateWidth - Default = 12
- ExpressionIntegerWidth - Default = 11
- CerlogFileSizeThresholdMB - Default = 0
Cerberus.INI configuration setting to control the growth of the Cerlog allowing definition of an integer value indicating the number of megabytes (MB) the Cerlog can grow to before Engine attempts to rename it and start a new file. If set to a non-zero value Engine will allow the Cerlog to grow to the specified number of megabytes before attempting to rename it. At the point the size of the Cerlog reaches the threshold Engine will flush pending log entries and rename the file using the file creation date to create a new name. The creation date is the UTC file time and should line-up closely with the first entries in the log file.
|