PDF of this tutorialSeveral industry trends are causing network equipment providers to move toward a more standard, less proprietary means of operation. Ultimately, technology investments must be flexible enough to accommodate new services incrementally, which ultimately can reduce the time to market and the cost of implementing these new services. To incorporate this flexibility, providers must shift from proprietary systems toward open, industry-standard software.
Conventional relational database and object-oriented database management systems provide a great deal of the functionality needed to power voice and data networks. As commercial off-the-shelf products, they inherently deliver tremendous time savings when compared to an organization building its own database.
Furthermore, these databases have a well-established history of excelling at many of the data management tasks required—such as managing customer data and automating on-line transactions. Additionally, a wealth of proprietary and third-party application-development tools are available that streamline application development for these databases.
However, these data management systems are typically disk optimized. Consequently, their performance is hampered by their fundamental architectural emphasis on optimizing the management of data on disk. Although conventional disk-based data managers increasingly utilize memory pools in their architectures to enhance transaction performance, the benefits of these steps are diluted by the extra management of multiple data locations, disk-oriented data structures, and disk-weighted optimization algorithms. As a result, these systems run into barriers that prohibit them from delivering the real-time responsiveness required.
