A service provider can make a difference for enterprise IT organizations by supporting and enhancing the company's business continuity plans to keep important functions operating acceptably when the primary IT infrastructure fails.
Service provider communications and hosting services can play a vital role for IT and its disaster/recovery initiatives. This requires redundant systems, most likely in two or more locations. IT's economic goal is to keep the costs low, but redundancy requires extra IT and communications facilities.
According to the Strategic Research Corp., a market and business development consultancy, the primary failures encountered in IT are:
- Hardware (44%)
- Human error (32%)
- Software and firmware errors (14%)
- Security breaches and viruses (7%)
- Natural disasters (3%)
Determining the reason for the failure is important, but recovery is the real objective. The IT organization has to consider the time it takes to recover and how much information needs to be recovered. Non-critical applications can be deferred until repair has been accomplished, while critical applications should have as little interruption as possible. This is exactly the point where service providers can help mitigate problems and offer cost-effective solutions.
Where service providers come in
In Forrester Research's report, IT Consolidation Drives Active-Active Data Center Configurations, one of the survey questions asked enterprises whether their internal data centers acted as recovery sites for each other. Seventy-four percent of respondents said they have an internal backup site; 24% use a service provider's disaster recovery service, and 4% have no recovery site.
Service providers can increase that 24% through customer education into the value of sharing disaster/recovery facilities at a service provider site rather than going it alone within the enterprise.
Data centers and server farms are centralized resources of computing and storage, as well as expensive real estate. Expanding the data center and server farm to accommodate new applications and IT communications adds stress to the operations and can be costly. Enterprise data center and server farm managers are always reluctant to keep adding servers or add new remote sites or another data center.
A service provider can reduce this stress by hosting new or moving non-critical applications to the hosting site on a permanent basis.
How to weigh the location of operations
The economic drivers that influence IT cover a wide range of factors. There are obviously hardware and software costs, and IT can do nothing without staff. New applications and constant changes add to the cost of IT operations.
Like the swing of a pendulum, most CIOs have been through variations of centralized, decentralized, distributed, hierarchical and peer-to-peer configurations. IP technology and related applications can be located in one or many locations and configurations. The cost to support and operate one or more sites will depend on several factors.
One location may be more vulnerable to an outage. A centralized location is easier to manage and staff. Remotely located staff diffuses the organization's expertise and limits the ability to respond the unanticipated situations and problems. Sometimes, decentralized IT sites will be required to meet regulations imposed by local or national jurisdictions.
The service provider can eliminate the cost of running a second data center in a decentralized enterprise data center model by hosting the remote site for the enterprise. This will allow the IT organization to avoid the disadvantages of decentralized operations (below).
Decentralized IT operations
- Shared load
- Less vulnerable to total failure
- Meet national regulations
- More staff
- Multiple locations
- Increased air conditioning
- Complex inter-site synchronization
Centralized IT operations
- Smaller staff requirements
- Efficient use of floor space
- More efficient air conditioning
- Less complex interconnections
- Site failure = total failure
- Inability to meet some foreign or domestic regulations
The idea of a backup location is now common for IT operations. Many were originally designed to take over the full traffic load and stayed in standby mode called an active-passive configuration. As the economics of IT became a bigger issue, the utilization of the backup site for live traffic has grown, according to the Forrester report.
The Forrester report stated, "Historically, alternate data centers have been an expensive insurance policy; the duplicate resources in this data center (DC) stood idle so that in case a disaster or some other calamity disrupted the production data center, IT operations could failover production workloads to this site to minimize downtime and keep the business running. But now, the pursuit of IT consolidation and greater financial and operational benefits is driving US enterprises toward more active-active data centers — data centers that run production workloads but also serve as recovery sites."
What this means is that all active sites must have equal communications capacity and connections. A majority of the time each site will support half the load and will be underutilized. Further, the active sites must be interconnected with sufficient bandwidth so the multiple sites can be synchronized within milliseconds of each other. The inter-site communications must have high bandwidth with low utilization, less than 5%.
So instead of the enterprise supporting two active-active sites, service providers can position themselves as the other active site for the enterprise.
Gary Audin, president of Delphi, Inc., has more than 40 years of computer, communications and security experience. He has planned, designed, specified, implemented and operated data, LAN and telephone networks.