Main Navigation (Explore Website):
Data Warehousing
We have long been concerned that the data warehousing market is treated as if it was homogeneous, with all vendors potentially tackling all warehousing issues. This is very far from the truth: for example, some products are specifically focused on supporting analytics while others are targeted at enterprise data warehouses. Similarly, some suppliers target the low end of the market, measured in gigabytes or perhaps a few terabytes, while others concentrate on implementations with tens or hundreds of terabytes, or even petabytes. Trying to compare all of these offerings within a single paradigm makes no sense. We have therefore elected to break the market down in terms of both functionality (analytic marts and warehouses on the one hand, and enterprise data warehouses on the other) and, in the case of the former: scale (small, medium and large). We have not distinguished in scale terms across enterprise data warehouses because the two tend to go hand in hand. Each of these four sub-market examinations is available as an individual paper in its own right, downloadable from the links below, but each refers back to the Market Guide as its foundation document in which we discuss market requirements and vendors.
There are around 20 vendor’s products to evaluate. However, it should be clear that they do not all compete with one another: there are multiple sub-markets within the data warehousing domain and, in addition to general requirements, we will also need to consider the particular markets that the various products are targeted at. However, as a general principle we are concerned with anything involving substantial amounts of (complex and ad hoc) analytic processing as opposed to environments where query and report processing is well-defined in advance. For requirements in the latter category, standard merchant or open source databases should prove sufficient for most purposes, though there may still be advantages in moving to one of the more specialised products. However, our focus is on environments that include heavy duty processing of data, which is why we have named the foundation paper "Analytic Warehousing".
Analytic Warehousing
By: Philip Howard
In this paper we have attempted to give an overview of the main concerns in the market and those vendors that currently play in it.
Small-scale analytic data marts and warehouses
By: Philip Howard
This paper covers small-scale analytic data marts and warehouses. That is, environments where there is a substantial requirement for complex and/or unpredictable queries but not things like master data management or operational BI (or only to a limited extent).
Medium-scale analytic data marts and warehouses
By: Philip Howard
This paper covers medium-scale analytic data marts and warehouses. That is, environments where there is also a substantial requirement for complex and/or unpredictable queries but not things like master data management or operational BI (or only to a limited extent). Note that not all of the vendors included in this paper have the ability to scale all the way up to 50Tb.
Large-scale analytic data marts and warehouses
By: Philip Howard
This paper covers large-scale analytic data marts and warehouses. A number of vendors that appeared in the two smaller scale analytic papers do not appear in this update. In some cases this is because they do not physically have the ability to scale to 50Tb but in others it may simply be that they have not demonstrated that capability even if they have the theoretical ability to do so.
Enterprise Data Warehouses
By: Philip Howard
This paper discusses enterprise data warehouses. By this we mean an analytic warehouse that also incorporates requirements to support such things as master data management and operational BI where there are a lot of look-ups and small queries running alongside the more complex needs of analytics. This puts a much greater emphasis on the ability to manage mixed query workloads. As a result, we have excluded from this update all products that do meet a certain level of functionality with respect to mixed queries.