Whether businesses operate out of single location or have multiple branch offices, they are increasingly dependent upon software applications to run their operations. If the applications go down, manual processes are often inadequate to keep the business going. The challenge is particularly difficult in the branch office, where there is almost never an on-site IT resource to deal with problems as they happen. To deal with that, companies have a few options:
- Install multiple servers and cluster them so that if one server fails, the other one can take over.
- Use applications that are provided from some centralized or regional data center or a 3rd party software hosting provider that can provide great application availability, and then access the applications using terminals or PCs over a network and hope that the network doesn't fail.
- Take a hybrid approach, with some applications hosted at a centralized data center and some operated in the branch on in-branch computers.
For applications that are hosted on computers in the branches, hypervisors from Citrix, Microsoft, or VMware, can provide the server clustering capabilities and help avoid in-branch server sprawl, while remote-management software, sometimes offered with the hypervisor, can eliminate the need for on-site IT staff. Depending upon the supplier and the desire for remote management, the branch will require at least two servers and perhaps a third for management. Failover capabilities also require shared storage, which means either installing a SAN, NAS, or using a storage virtual appliance running under the hypervisor. Combined, these dramatically reduce the probability that applications in branch offices will go down.
There are many other benefits to this approach, including being able to eliminate emergency and off-peak service calls, being able to install software updates without taking applications off line, and requiring no on-site or regional-area technical staff to manage the servers and applications.
For software applications that are housed in corporate IT or delivered from third parties as a service, organizations need to consider the impact of network downtime on business operations. Organizations can utilize redundant networks from different suppliers, or use dial-up networks as a degraded-service-level alternative in emergencies. While costs for networks have declined significantly, multiple networks may be prohibitively expensive for some operations. Organizations also need to consider which processes can continue in a manual mode, if the centralized or 3rd-party hosted application fails.
One alternative to this network dependency for hosted applications is to also make available a subset of applications and data in each branch. In this way, and within limits, the business can continue until the hosted application is restored.
Action Item: Before deciding on an approach (in the branch, in the cloud, or a hybrid approach), companies need to evaluate:
- Which applications are critical to the operation of the business and have to be up and running every day.
- Which applications can be down for a day or more, without impacting the business.
- Which applications can being delivered as a service over terminals and a network.
- For applications that are being delivered as a service, what hybrid options are available to continue to operate when the hosted application or the network goes down?