Description
LAN High Availability Configuration Solution is for expected uptime of 99.999 percent or higher. At the device level, neighboring user devices are connected to alternate workgroup switches so that a switch outage will still allow some users to work unaffected or other users to use unaffected nearby machines. Redundant server switches allow for redundant connections to the servers so that if one switch goes down, the server is still accessible without any remedial action required. This pattern also includes the addition of redundant access and distribution switches and redundant paths between them. The workgroup and server switches are dual-homed to the access switch. The distribution routers are dual-homed to core routers. Because the distribution and core routers are dual-homed, the failure of an access switch will not impact availability. The LAN High-Availability Configuration Solution can also include load balancing functionality, which is described in the Network Load Balancing patterns.
The high-availability configuration option should be used in mission-critical subnets, locations and data centers where uptime is critical to the mission of the NIH. For example, the Clinical Center provides healthcare to patients in research trials, and the impact of network connectivity downtime could be life-threatening for patient care; therefore, the High-Availability LAN configuration is an appropriate option to support this environment. Grants management is another mission-critical area for NIH. Since grant awards are often time-sensitive, especially around the end of the fiscal year, some subnets that support grants evaluation and notification processing may warrant the High-Availability LAN configuration solution. Calculated availability service levels (such as ?99.5% or 99.999%) are predicted based on statistical probability and historical outage experiences of selected products and other factors, and are not guaranteed. It is important to work with the user community within the context of overall continuity of operations to understand when standalone operations will suffice as a backup method in the event that the cost of high-availability options may not be justifiable.
Diagram
Benefits
- Provides a consistent and manageable approach for provisioning LAN services across the NIH. Can enable a breadth of configurations that are neither under- nor over-configured.
- Allows the alignment of technical requirements to business needs by implementing LAN service levels based on availability requirements. For example, if only 99.5 percent uptime is required for a specific location, a design pattern can be developed for that requirement. Because it has interchangeable layers, this design pattern can be easily customizable for varying uptime requirements.
- Provides an efficient use of expensive network resources, such as printers, scanners, and other network devices.
- Provides the ability to aggregate large numbers of users onto expensive WAN transport services.
- Provides a greater opportunity to avoid outages that might impair the NIH mission or cause damage to its reputation through unplanned outages.
Limitations
- Implementation is more expensive than the minimum configuration pattern.
Time Table
This architecture definition approved on:
February 8, 2005
The next review is scheduled in:
TBD
