It’s been a tough week for networking in the enterprise. Key outages brought large sectors of the economy to a standstill, and while the causes were miniscule, the effects were not.
Most of the attention went to United Airlines, which had to ground more than 800 flights due to problems in their ticketing and registration systems caused by a faulty router. Engineers bypassed the problem within a few hours, but by then, the ripple effects had spread to connecting flights, gate reservations, and a host of other scheduled events.
A short time later, the New York Stock Exchange (NYSE) went dark for nearly four hours due to an error upgrading the timestamp software in its electronic trading infrastructure. Simultaneously, the Wall Street Journal went offline due to what some reports called a problem with its web gateway.
These problems illustrate the difficulty that networking presents to large organizations as they transition to the all-digital economy, says Computerworld’s Stephen Lawson. The fact is, most large organizations utilize redundant network resources, automated failover and a host of other tools designed to circumvent total network meltdowns, but remain vulnerable to minor problems on individual devices that then produce a series of cascading issues that end up hampering critical applications and services. Typically, these problems arise during software upgrades and systems integration processes, as was the case earlier in the decade, when United saw several system failures following its merger with Continental.
Software defined networking (SDN) would fix these problems, according to CRN’s Mark Haranas, but only if it is properly configured and utilized. The problem is, no one has attempted to push SDN onto a wide-scale, customer-facing data portal like a flight reservation and booking system, so it is still rather difficult to say how immune it will be to software issues or mismanagement. One thing is clear, however. As data loads increase, simply adding to the same old network infrastructure will not be enough. Emerging data platforms require a network that is more nimble than bulky. For that reason, incorporation of fabric-style topologies, preferably on the abstract, virtual plane that is the hallmark of SDN, will be the only way to handle the dynamic nature of the modern digital economy.
And even though the issues that led to the United and NYSE outages were not that serious and the fallout was relatively mild compared to previous disruptions, they raise troubling questions about the growing dependency on digital networks in modern society. Computer Weekly’s Warwick Ashford notes that if these issues were to take down key national infrastructure networks, like an electrical grid, the consequences could be catastrophic. With so many layers of digital technology now standing between consumers and the power we use, the food we eat, and the way we communicate, the chances of a consequential malfunction are increasing.
This is why we should not try to harden our networks against failure as much as plan for speedy recovery when they do go down, says the LA Times’s Andrea Chang and Tracey Lien. Technology can never be foolproof – fools are simply too ingenious – but as the world becomes wholly dependent upon the digital ecosystem for virtually everything we do, the organizations that build and manage this infrastructure need to spend more to ensure security and reliability. Currently, only about 10 percent of the IT budget goes to security, says TrapX general manager Carl Wright.
The lesson to be learned from all this is not that we should start panicking over the safety of our data infrastructure. Instead, we should recognize that there are vulnerabilities now, and there will continue to be vulnerabilities in the future. A more proactive approach to network management will go a long way toward preventing serious disruptions, while greater contingency planning will help lessen the real-world impact when the bits stop flowing.
Photo courtesy of Shutterstock.