Inside Data 64
by Graham Keitch
Graham Keitch checks out your options when it comes to databases in the cloud.
HardCopy Issue: 64 | Published: October 30, 2014
Web technologies have changed the data landscape beyond recognition, resulting in the evolution of more appropriate models and infrastructures to help us work with the data. The web has accelerated our expectations for mobile cross-platform data access too, forcing IT departments into uncomfortable territories. The cloud can help businesses extend data access and add new technologies to their on-premise facilities, often with lower risk and set up costs. In some cases it may be the only viable option. For example, let’s consider how database technology is responding to our requirements when it comes to processing and analysing unstructured data.
Conventional relational database management systems have been refined over decades to provide optimal performance for online transactional processing (OLTP) and are not best suited for analytical work. Analytics is executed more efficiently when data is stored in a columnar format. Microsoft SQL Server 2014 has new features to support columnar data, while Oracle Database 12c combines column store with in-memory technology.
Different models are required for unstructured data, which has led to the creation of the NoSQL (Not only SQL) database and mapping techniques for data navigation. The business intelligence (BI) tools provided with Microsoft SQL Server 2014 integrate with the Hadoop-based HDInsight for analytics on unstructured data. Oracle has a good track record for supporting unstructured data types such as DICOM medical images, and it also has its own NoSQL database offering.
There are two approaches to deploying a database in the cloud. The first involves a virtual machine that allows the user to upload their own database image, or use one that has been provided already. Examples of relational SQL databases include Oracle Database, Oracle MySQL and IBM DB2 while non-relational requirements are catered for by Hadoop, Apache Cassandra and MongoDB via Amazon EC2, Microsoft Azure or Rackspace.
The alternative database as a service (DBaaS) model frees the user from installing and maintaining either the database software or hardware as this is handled by the service provider. Otherwise, running your own database in the cloud is similar to running it on-premise. Microsoft’s SQL Azure is well established in this space to serve traditional RDBMS requirements, while various offerings from Amazon, MongoDB and others exist for NoSQL.
DBaaS provides an opportunity for the vendor to offer tailored services such as Oracle Data as a Service for Marketing, launched July 2014 as part of the Oracle Data Cloud solution. DBaaS can also be useful for systems that might not be so easy to ramp up on-premise, especially where cross platform and other complex issues are being tackled.
The cloud can help extend applications for mobile platforms. Azure allows you to build and host the backend for mobile apps across numerous platforms including iOS, Android, Windows and HTML5, as well as facilitating integration with SQL Server, Oracle, SAP, MongoDB and others.
For many organisations, the priority centres on making existing data assets available to cross platform devices using internet technology, but not necessarily a cloud database. Oracle’s Mobile Suite uses Java and HTML5 for the logic and client side interface, allowing you to build applications for iOS and Android. The applications connect to backend services using both REST and SOAP interfaces. Oracle Mobile Server 11g provides the synchronisation and management features required to connect large numbers of users to the backend Oracle Database.
SAP Sybase SQL Anywhere is an established cross platform solution for running and synchronising mobile applications with minimal administration. It can be used as a workgroup server and a mobile database with change-based replication between backend systems such as Oracle Database, SQL Server and Sybase’s own database family.
The cloud provides database developers and administrators with an environment for prototyping and offsite storage. Microsoft Azure offers various storage options in the cloud including Blob Structure for unstructured data. The cloud also makes it easy to integrate with other services such as Microsoft’s Bing Maps.