The Reality of Real-Time Vaulting
- Published on Friday, 26 October 2007 16:35
The preceding paragraph highlights the process of “electronic vaulting”, that is, extending the capabilities of a data center to an alternate site (vault). As described, one or more channels of the production processor are directly connected to the alternate site, where tape subsystems are dedicated for normal backup data protection operations.
Channel extension technology is readily available that replicates standard bus/tag protocol in such a manner as to provide a transparent interface to remotely located devices. The actual electrical connection is provided by high-speed carrier services including T1, T3, or private (dark) fiber installation. Use of a T3 service for example, can enable subsystems to operate in data streaming mode, at maximum channel speed (3 megabytes per second), over great distances.
What are the Benefits?
Armed with the knowledge of how the techonolgy works, potential users of electronic vaulting services must evaluate the key benefits to be derived from this latest service offering. Unlike former methodologies that rely upon traditional approaches to offsite data storage (local dumps, staging, media shipment), electronic vaulting users can take advantage of direct transmission of critical backups to the vault. This method reduces or eliminates the physical movement of media, while effectively addressing corollary issues such as risk of loss during transport, creation of multiple backups of the same data, logistics and administration, etc.
The immediate nature of electronic vaulting assures target data has been secured off-site without needless exposure to loss during the lag time that the media awaits offsite shipment. Channel extension transparency implies that all current data backup and restore software operates unchanged when used for vaulting operations, including volume backup systems, database image copies, user-written storage utilities, active and archive database logging programs, incremental backup software, as well migration volume support employed by system managed storage-oriented subsystems. In addition, local production recovery operations are significantly expedited. Database corruption, inadvertent or deliberate data erasure, and so forth, are situations that mandate immediate access to archived data for the purpose of local critical system recovery. Former approaches relied once again upon the physical selection and transport of the correct volumes from the offsite storage provider to the production data center, with the corresponding time delay. Electronic vaulting technology eliminates delays associated with production related system recoveries. Because all archived data is effectively online, various recovery programs simply call for the correct dataset names and volume serial numbers are immediatley loaded on the vault-located subsystems and restored as required.
Another direct benefit is the assurance of critical transaction survival. Certain datasets, referred to as journals or logs, reflect database update activity since the last backup cycle. Whether the user’s requirement involves CICS journaling, the maintenance of active logs for database systems such as DB2, IMS, IDMS, etc., or the employment of batch program journals (if the right software is installed), critical databases can be recovered up to the point of unavailability or corruption. Overall, the entire process is straightforward: the user loads or restores the last backup of the data and utilizes a recovery utility to apply all update transactions recorded in the journal (log) file. The end result is that data integrity has been preserved and users can pick right up where they left off without the need to re-enter transactions and are assured that critical transactions entered throughout the production day are recovered.
Basically, this scenario works out fine for local recovery where data center operation is usable, but what if the database reconstruction has to occur at the disaster recovery center due to an unplanned event? Electronic vaulting supports the continual or periodic storage of active and archive journal logs at the vault, thereby ensuring that if a disaster occurs, business activity can resume starting at point of failure without loss of critical transactions. Availability of these transactions can have significant impact on bottom-line corporate profitability.
In addition to data integrity, electronic vaulting improves contingency plan testing because major tasks associated with the periodic testing of contingency plans at an alternate site or recovery center includes defining, assembling, and shipping backup media. With the vault located at the recovery site, such logistical issues are diminished for the most part.
Also, tests can be executed more frequently, can use “fresher” data, and can be accomplished with a considerable reduction in expense. Perhaps the most significant benefit of electronic vaulting is the dramatic reduction of time required to restore the production system at the disaster recovery site. It also avoids the possibility of data loss during a disaster due to either misplaced or damaged media during shipment. Because all of the corporation’s vital data is located at the disaster recovery center, the staff employed at that recovery facility can initiate system restoration well in advance of the user’s arrival.
Electronic vaulting checklist
When evaluating electronic vaulting services, a comprehensive checklist should include a number of items in the areas of security, capability, performance, and disaster recovery availability.
First, the vault location should exist at the site where the user plans to recover operations during a disaster. If this involves a commercial site, it must be self-contained, environmentally conditioned, secured, and accessible only to the vault staff (supplied as part of the service). It is absolutely imperative that no equipment at a vaulting facility be shared; every user must have dedicated capability.
The vault should be of sufficient size to contain much of your total offsite media needs and must accommodate a variety of channel extended equipment, including tape subsystems, operations consoles, printers, etc. A portion of the justification strategy might be utilization of excess vault space for special equipment needs (link attached 3725, production network nodes, etc.) not related to electronic vaulting.
Furthermore, electronic vaulting must use channel extension hardware which is totally transparent to both systems and applications and does not impact the production network. It is critical that the user carefully evaluates (and avoids if possible) installation of special software required to make the vault operational. Also, the user must validate that all of the existing data management software and database systems operate in conjunction with the vault without modification or reformatting.
If the user’s primary interest is update transaction capture at the vault, it’s important not to limit the operation to a single database system such as IMS or a single online application such as CICS. Another important consideration is the batch world and the user shouuld investigate currently available systems which have the ability of journaling batch updates and reconstructing standard VSAM files, in a manner analagous to the logging and recovery utilities available to online systems.
Finally, the user must evaluate performance and expandability of the communications circuits connecting the production complex to the vault. The channel extension hardware should support T3 or private fiber to maximize capability, and support low-speed (T1) carrier for subset operations such as critical transaction capture. Overall, users must take a look at the everyday production operation and do a quick mental audit of the recovery capability focusing on availability and currency of critical data. If some of the benefits of electronic vaulting apply to specific user organizations, the implementation can be readily accomplished by an existing staff or through a provider of commercial disaster recovery services.
This article adapted from Vol.1 No.3, p. 16.