| The method data warehousing vendors use to | | | | projects and the level of confidence |
| provide their service is a key issue. Vendors who let | | | | Constraints on monetary resources, cost and time |
| you run your own server's license or those who host | | | | benefits derived from implementing a solution |
| software tools on their servers are common options. | | | | compatible with existing systems |
| There are multiple options related to the architecture | | | | The need for a point solution for a particular business |
| of the data warehousing system. The most popular is | | | | unit, or a decision support infrastructure which can |
| the hub-and-spoke architecture, which is a centralized | | | | support a range of applications |
| data warehouse with dependent data marts. It is a | | | | The ability to integrate metadata, scalability in terms |
| sort of corporate information factory. | | | | of number of users, volume of data and query |
| One other choice in data warehousing architecture is | | | | performance, ability to maintain historical data and |
| the data-mart bus architecture linking dimensional data | | | | adapt to changes related to source systems |
| marts, or a centralized data warehouse with no | | | | Depending on your organization, these factors may |
| dependent marts and last but not least, a federated | | | | be considered to have a lower or higher priority. A |
| architecture. | | | | lack of resources may prevent some organizations to |
| Some organizations develop their own data marts, | | | | implement a better architectural data warehousing |
| which are independent from one another. However, | | | | solution. The centralized data warehouse architecture |
| they have inconsistent data definitions and different | | | | is faster and easier to implement since it does not |
| dimensions, which make it difficult to analyze data | | | | require dependent data marts. |
| across marts. | | | | The design of user interface has an important impact |
| Data mart bus architecture implies building a first mart | | | | on the decision to choose a certain vendor. A tool |
| that uses dimensions and measures that will be used | | | | that is uncluttered, intuitive will be better used by |
| with the other marts that are developed, in order to | | | | intended users. Data warehousing solutions assume |
| obtain logically integrated marts. The data is organized | | | | that users are trained in data management and |
| in a star schema and this provides a dimensional view | | | | analysis techniques. |
| of the data. | | | | The support offered by data warehouses comes in |
| The hub-and-spoke architecture focuses on building a | | | | two different forms: technical and functional. While |
| scalable and maintainable infrastructure. It is | | | | technical support focuses on hardware, application of |
| developed in iterative manner. Dependent data marts | | | | software patches and updates, functional support |
| obtain data from the warehouse and they can be | | | | includes user manuals, training sessions, and client |
| developed for different departments or special | | | | services. |
| purposes. | | | | Security of the data warehousing services refers to |
| A similar data warehousing architecture is used for | | | | the physical and electronic aspects. Physical security |
| centralized data warehouses but without dependent | | | | involves restricted access to the servers which store |
| data marts. Queries and applications access data | | | | data and off-site emergency backups. Electronic |
| from relational data and dimensional views. | | | | security involves limiting access to data and use of |
| Federated architecture allows data to be accessed | | | | software to identify unauthorized attempts to |
| from sources such as operational systems, data | | | | access the data. |
| marts and data warehouses. It is logically or physically | | | | The legislation regarding data ownerships is not |
| integrated by use of shared keys, global metadata, | | | | keeping the pace with the advance in technology and |
| and distributed queries. It is an adequate solution for | | | | business practices. You have to take this into |
| companies that have a complex decision support | | | | consideration and spend time on analyzing any issues |
| environment. | | | | that can arise from this incompatibility. |
| When deciding on a data warehousing solution, you | | | | The reputation and history of vendors of data |
| may need to assess the following advantages of | | | | warehousing services should be investigated |
| different architectures: | | | | thouroughly. References and people in the industry |
| The level of information interdependence and the | | | | can be a valuable source to assess the quality of |
| ability to share and integrate information | | | | customer service offered and the integrity of the |
| Senior management needs regarding information from | | | | vendor. A visit to the physical space is also helpful in |
| lower levels | | | | assessing the quality of the organization. |
| Urgency of the need for capabilities of data | | | | Once the selection is made, the exact services to be |
| warehousing architecture | | | | provided, expectations related to customer service, |
| Non-routine tasks, the need to analyze data in novel | | | | contact information and payment will take priority. |
| ways | | | | You will benefit from a better return on investment |
| The availability of IT personnel, the technical skills of | | | | and higher quality of service once you reach an |
| IT internal staff, successful experiences with similar | | | | informed decision. |