You've Got the Bandwidth, but How About Availability?
All the bandwidth in the world won't mean a thing if you lose availability. Invest in measures to prevent costly network downtime—it's worth it.
Last week, I talked about broadband capacity on the global network infrastructure, which I concluded turns out to be more than adequate for the data needs of today and the foreseeable future.
Within the enterprise, however, the threat of overburdening available network resources takes a back seat to the fear that networks will suddenly become unavailable through either system failure or management mishap. All the bandwidth in the world (literally) isn’t worth a hill of beans if connectivity is lost.
Ultimately, availability is a matter of cost. According to Gartner, network downtime costs North American businesses close to $26 billion a year, derived primarily from the estimated 127 million man-hours lost as well as indirect costs like the erosion of customer and partner confidence and growing employee mistrust of IT. Retailers in particular are highly sensitive to network disruptions, particularly around holiday time (natch).
And the fact is that without some remedial action, enterprises will see network availability drop rather than rise, even as new virtual and software defined technologies are deployed. As I mentioned in my last blog post, bandwidth both in and out of the datacenter is in ample supply, but with traffic loads increasing thanks to Big Data and the flood of mobile users, expect sudden spikes to hit enterprise networks with increasing frequency, potentially overloading available capacity. As the U.K.’s bdaily.com reported recently, an English steel plant recently suffered a near-crash because one employee, one, was downloading training films to a personal cell phone.
In most cases, proper data policies can blunt the impact of rampant data activity on enterprise networks. But new data infrastructure designs hitting the channel may also improve availability by extending system redundancy across server, storage and network architectures. A prime example is the Cisco/NetAppFlexPod, which was recently upgraded with the Nexus 7000 switch in support of high scalability and high availability for 24-hour operations. The system now provides up to 768 10 GbE ports and end-to-end Fibre Channel over Ethernet connectivity that can be used to build unified Ethernet fabric architectures across multi-data center deployments.
In addition, many organizations are turning to outsourcing and Network-as-a-Service solutions to gain the kind of redundancy and automated failover needed for always-on connectivity without blowing the entire capital budget. MegaPath Corp., for one, has launched an “any-to-any” Broadband Managed Failover service, designed to kick in if primary data connections fail. The service supports Ethernet, T1, DSL, and cable networking, with a companion wireless service available for 3G, 4G, and LTE backup, and features an on-premises active monitoring component that automatically detects a disruption and activates the failover. The company also provides a repair and restoration service that automates the recovery of primary circuits and returns connectivity to its original state.
Nobody likes a loss of service, and in our connected world, people get nervous when their connection fails for even a little while. Disruption of the enterprise network, then, is more serious than loss of server or storage resources. It can bring all productivity to a standstill. As such, it deserves a significant portion of both capital and operational budgets.
After all, the cost of maintaining availability is a pittance compared to the cost of losing your network.