In 1992, the SHARE user’s group defined seven levels of disaster recovery capabilities. These seven levels have been widely embraced by the industry and are commonly used by MIS organizations, vendors, and consultants for planning and discussing recovery capabilities.
These seven (actually eight and sometimes nine) levels are frequently referred to as DR “tiers.” However, the numbering of these DR levels is reversed from the traditional numbering of MIS service level tiers. Level 1 has the lowest DR capability, as well as the lowest cost and complexity. Levels 6 and 7 have the highest DR capabilities, as well as the highest costs.
The seven levels assume that MIS operations take place in a primary data center, with the availability of an alternate data center for use in a disaster. The seven levels concentrate on how data and how application transactions are transferred from the primary data center to the alternate data center. The seven levels describe the technologies and investments needed to provide faster recovery following a disaster. Recovery times are conventionally measured as Recovery Time Objectives (RTO) and Recovery Point Objectives (RPO.)
One of the most important concepts of the Seven Levels is that a specific desired recovery level must be designed into the specific architecture of each application. Achieving faster recovery times requires designing applications and infrastructures using specific technologies and methodologies.
One of the greatest recurring problems in Business Continuity planning is disconnection between DR expectations and the level of investment in DR designs and infrastructures. It is extrememly common for CEOs, CFOs, and business unit managers to have expectations for fast recoveries (usually corresponding to DR Levels 3-5,) while actual infrastructures have been implemented at DR Level 1 based on the available DR budgets.
As presently defined, the seven levels are tied to specific server and storage technologies and recovery methodologies, particularly technologies and methodologies popular in the 1990s and early 2000s. There have been some important new technologies (particularly server and storage virtualization) which are not directly addressed in the seven levels. It is possible that, in the future, the definitions of the existing levels will be more generalized, or there will be a few additional levels defined.
As seen in the following diagram, moving to a higher level provides faster recovery times, but requires a significantly higher investment in hardware, software, telecommunications, facilities, training, services, and management overhead.
The seven tiers have been described as follows:
Tier 0 – No data is preserved off-site. No recovery is expected.
Tier 1 – Data is preserved off-site in backup format.
Tier 2 – Data is preserved off-site in backup format, with a DR hot site available.
Tier 3 – Data is preserved off-site in backup format using electronic vaulting.
Tier 4 – Data is duplicated off-site as point-in-time copies of primary data.
Tier 5 – Applications are customized for transactional integrity across two (or more) data centers.
Tier 6 – Data is duplicated off-site as continuous copy of primary data.
Tier 7 – Data and applications are duplicated off-site, with automated recovery.
Meeting specific recovery requirements is heavily dependent on leadership, organization, documentation, staff skills, and practice exercises. At the end of the day, only DR testing will ensure that any enterprise and infrastructure is capable of meeting specific recovery requirements.