Proxy based approach for space optimisation

3.5.1 Proxy based approach for space optimisation

The rising of data volumes in many application domains, makes raise a problem to maintain such large data storages. Though, the storage space is reducible if the space of storage was optimised. These space optimisations not only contribute to save the space, but also decreasing the carbon footprint and the cost of operation. Beside this, it also optimises query response time. Additionally, this space optimisation might make easy the job of administering which are basically requires new infrastructure, utilities like power and cooling increase, extra floor space and extra staff. Emran, et al., 2012. Therefore, Emran, Abdullah, and Isa 2012 produced an approach called Proxy- based approach which can generate space optimisation through modification of database schema. This can be done by deleting the attributes from the particular schema. The term ‘proxies’ were used by the researchers, is to replace the attribute with another attribute in the schema. The functional dependency relationship is used to recognise the proxies among the attributes in a relational table. Basically, the space saving is obtained through some modification in the schema by dropping some of the attributes. And then, the total saved space is approximately verified by the number of attributes has been dropped and the tuples number in the table remain. However, the droppable attribute and the proxy must have relationship in terms of missing data. Hence, the functional dependency relationship is obtained between the attributes in the relational tables. Proxies for the delete able attributes been found through discover of the relations among attributes in the tables where there is presence of FD. This proxy-based technique apply algorithm which will get back the removed overlaps from the meta-data, when the query is submitted against the compressed tables. 32 For instance, a and b are the droppable attribute. A proxy map consists of the following mappings: a → {1,2,3,4}, b → {5,6,7,8}, where the numbers are the proxy values and the arrow shows relationship mapping. From here, the researchers have identified two types of proxy maps as follows: i. A pure relational table: This structure shows each value in a droppable attribute is matched to exactly one value of the proxy. The schema structure of the table is: droppableAttr, proxy. Table 4 ii. A multi-valued table: In this table, each value of droppable attribute is matched to a set of proxy values. The schema structure of the table is: droppableAttr, proxy. Table 5 Table 4. A Proxy map in pure relational table Emran, Abdullah, and Isa 2012 A B a 1 a 2 a 3 a 4 b 5 b 6 b 7 b 8 Table 5. A Proxy map in a multi-valued table Emran, Abdullah, and Isa 2012 33 A B a 1,2,3,4 b 5,6,7,8 The example also shows that, the size of proxy map in the multi-valued table is smaller than the proxy map in pure relational table. As a result from the example, the storage space can be saved by minimizing the proxy map in a multi-valued table structure as shown in table 2.2. In the example above, the multi-valued table saves of 6 instances.

3.6 Conclusions