Dependable Infrastructure and Data Management

Dependable data center infrastructure is more important than ever when operating in a complex and challenging environment where the demands for security, speed and reliability are tested with every transaction and interaction. Organizations are also realizing that well-managed data centers can make a significant impact to their bottom-line by enabling them to respond more quickly to market demands...

The Data Center Infrastructure Management (DCIM) professionals from CA Services have the expertise and skills to implement a unique solution that enables you to effectively manage your three biggest data center infrastructure management challenges: power and cooling, capacity and inventory, and IT and business facility resources.

Major components of dependable systems include dependable networks, dependable middleware including infrastructure, dependable operating systems, dependable data managers, and dependable applications. That is, we need end-to-end dependability. By dependability we mean security, integrity, fault tolerance, data quality, privacy, and real-time processing.

DEPENDABLE INFRASTRUCTURE:

For dependable systems, we need appropriate operating systems and infrastructure possibly based on object request brokers. We need to host sensor data managers and applications such as multisensor data integration and fusion on the infrastructure. We need to ensure that the system is secure and survivable. We also need to ensure that the infrastructure is secure. That is, security has to be built into the system and not considered as an afterthought. Essentially we are proposing a layered architecture for sensor information management. The infrastructure consists of middleware possibly.

Another aspect that is important is end-to-end dependability. This includes security, real-time

processing and fault tolerance for not only all of the components such as infrastructures, data managers, networks, applications and operating systems; we also need to ensure the dependability of the composition of the entire system. This will be a major challenge even if we consider security, real-time processing and fault tolerant computing individually. Integrating all of them will be a complex task and will have to address many research issues and challenges.

DEPENDABLE DATA MANAGEMENT:

Dependable data management systems are needed to process stream data. For example, much of the data may be transient data. Therefore, the system has to rapidly analyze the data, discard data that is not needed and store the necessary data. Special query processing strategies including query optimization techniques are needed for stream data management. Many of the queries on stream data are continuous queries. We need special query strategies for processing such continuous queries. Researchers are also examining special data models as well as extensions to the relational data model for stream data management. Query languages such as SQL (Structured Query Language) are being extended with constructs for querying stream data. Research efforts are also under way for extending XML (eXtensible Markup Language) with constructs for sensor data management. We also need to determine the types of metadata to collect as well as develop techniques to store and manage the metadata.

SECURITY ISSUES FOR DEPENDABLE INFRASTRUCTURE AND DATA MANAGEMENT:

•  These systems are vulnerable to unintentional errors: equipment failure, human errors, weather, and accidents
• They are also the object of intentional attacks: external (hackers) and insiders (internal attacks).
• We need to conduct research on security for sensor databases and sensor information systems.
• We need to examine security policies for sensor data.
• We need end-to-end security. That is, the infrastructures, data managers, applications, networks and operating systems have to be secure.
• We also need to investigate the composability of the dependable objects.