Increased demands on enterprise applications also come from other sources. Automated manufacturing processes, 24-hour service phone centers, ATM cash machines, and e-commerce demand service 24 hours a day. No longer is it possible to shut down a system every night to perform maintenance on operating systems and databases. Now a system must be available 24x7x365 to service the needs of a corporation’s customers and partners.
Today’s generation of intelligent enterprise applications are responsible for nearly every transaction that an organization has between its partners, clients and employees. The rapid adoption of collaborative technologies and applications such as XML, CRM, and private exchanges creates new, unforeseen challenges for backup, restoration, and recovery strategies.
Whether the application is a content, customer relationship, or supply chain management solution, the relationships between information within these systems (i.e., the metadata) are becoming more valuable – and mission-critical – than the data itself.
These relationships, when structured together (e.g., a workflow scenario), create an organization’s e-business processes; responsible for pushing and pulling content and data out to Web sites, suppliers, vendors, and customers. The relationships between documents, data files, image files, and other information, on an enterprise-wide scale, are growing extremely intricate and complex. Using parent-child relationships, static links, live links and automatically defined context-sensitive links, the possibilities for reuse of information in new and creative ways is endless.
The Importance Of The Relationships
It is essential to find a means to backup and recover these systems that have become a fundamental part of the foundation of intelligent enterprise applications. Integrity must go beyond traditional approaches (managing metadata via a RDBMS backup routine and documents via a file server backup utility).
Today’s applications use data stored within an RDBMS and documents stored within a file system. Some products exist for backing up an RDBMS. Other products exist to backup files and directories on a file system (including those marketed as enterprise solutions with hardware and software). The problem for the customer is which products to buy and how to integrate them into the existing environment.
Traditional backup solutions are designed to handle the data alone – a valuable and necessary component of the evolving backup and recovery strategy, but no longer the whole solution.
Restoring data without its relationships intact is like having data with amnesia.
The data doesn’t know to whom it belongs, where it’s going and what its relevance is to the business. Data without its relationships is not very valuable to the enterprise.
A typical backup and disaster recovery solution cannot preserve the business processes that are used to communicate with partners, employees, and customers. These processes provide the context for all of the data gathered.
Without the context and relationships preserved, there is no depth of understanding of what the significance of the data is and how it should be used in regards to a particular customer, partner, or employee. It’s like receiving a piece of paper with a name and phone number on it but with no further information. “Error! Reference source not found,” shows a typical object and its relationships with other objects within a system.
The knowledge contained within the metadata must be preserved by the backup and recovery solution. “Work in process” at all levels must be captured and synchronized. No longer is only the data at risk, but the knowledge captured in the form of links, ownership and process experiences.
A traditional, “full” backup does not capture this information. A “full” backup takes a “snapshot” of all the data in the system at a given time. This involves a complete backup of the RDBMS database via one product and a mutually exclusive backup of the file systems that store the content using a different product. A significant amount of work is involved in determining which products to use, the procedure to establish the backup and more importantly, how to restore the data.
Having to use unrelated products to backup metadata and content means more maintenance demands on the system’s administrator. These solutions make the task of individual object restoration tedious and time as well as resource (hardware and employee) consuming. Some of these solutions require restoring the complete image to another server and then locating the object.
In fact, the object may not even be in the backup if it was created before the last snapshot was taken. A backup method that backs up data and metadata is needed.
This method of operation allows individual objects to be restored; eliminating the need for a full restore if just an individual object needs to be restored (e.g., when a file is accidentally deleted).
Backup Need Not Mean Downtime
Today’s intelligent application environments must be available around-the-clock. The business continuity process cannot tolerate downtime, either to build backup files, or restore a system.
Recovery must be provided to any point in time, and done so in a complete and synchronized manner. A cold backup requires that the application be shut down. A backup that occurs in read-only mode is only a partial solution. This type of backup permits users to view data without having to shutdown either the application or RDBMS, while maintaining a consistent state between the metadata and content. However, this type of backup falls short of true 24x7 availability.
What is needed is a backup approach that operates continuously in background, while users continue to work. This has two benefits.
First, by operating in continuous background mode, recovery can be provided to any point in time. Recovery is not limited to the time of the last “scheduled backup.” Work that is performed between traditional backup periods is not lost if the system fails (subject to the capacity of the file system for backups and the threshold for data loss).
Second, the execution of the backup process does not infringe on the business process. Unlike traditional backup approaches and procedures, the backup continues to provide full read-write access to work objects, and execution of workflow scripts, during the backup operation. Thus, continuous backup is provided without negatively impacting critically important production time.
Conclusion
The goal of 24x7 availability required by today’s business environment calls for a new approach to backup and recovery of intelligent enterprise applications. The context and relationships among data must be considered critical information that must be preserved by the backup process. The backup process must be as unobtrusive as possible – today’s global business environment cannot tolerate scheduled downtime as it once could.
These goals can be approached through a combination of traditional backup methods and the use of a new generation of off-the-shelf backup and restore solutions.
Bruce L. Rudolph is chief technology officer and a co-founder of CYA Technologies Inc. Rudolph has more than 20 years experience developing software to support large technical and mission-critical applications for regulated industries and government agencies. His experience comprises all phases of the product/application development, including requirement analysis, specification design, implementation, and testing. Rudolph holds BS and MS degrees in Computer Science from Central Connecticut State College and Polytechnic University, respectively.
