Internal Cloud Bandwidth Concerns: The Buyer’s PerspectiveIntroLatency. Congestion. Extended delays.
To IT departments, users, and enterprise management alike, these are dirty words that represent frustration and annoyance. More importantly, they represent hits to productivity.
For enterprises moving internal applications to the cloud, it’s possible that sufficient bandwidth exists to prevent uttering these dirty words. For others, adding bandwidth and addressing network equipment shortcomings is a must, particularly when moving mission critical applications to the cloud and in scenarios where consistently quick access to time-sensitive data is imperative.
The trick, though, is determining how much bandwidth is enough and where those shortcomings might reside.
Hitting The Bandwidth Sweet Spot
Generally, IT departments overbuy or underbuy bandwidth because they lack a firm grasp on what the enterprise needs concerning applications, as well as what end-users’ expectations and demands are related to those applications. “It’s like insurance,” says Bob Laliberte, Enterprise Strategy Group senior analyst. Rather than be caught short, risk-adverse organizations take the approach that possessing excess bandwidth leaves them always prepared. Overbuying bandwidth, however, also occurs when IT throws bandwidth at a given problem rather than attempting optimization techniques.
Underbuying bandwidth can occur if IT lacks the budgeting and/or understanding to enable projecting future bandwidth requirements. Organizations can’t do much about budgeting deficits other than use historical and trending data to their advantage, Laliberte says. A lack of understanding, though, can indicate a lack of visibility into the existing environment. (Such an inability to use trending data to detect where more bandwidth might be applied would prove beneficial, for example.) Similarly, IT might base bandwidth projections on incomplete assessments. The latter often happens when IT adds storage replication apps but doesn’t account beforehand for the fact that they’re “huge bandwidth hogs,” Laliberte says.
There are cases where overbuying bandwidth is done intentionally. Michael Shonholz, CDW telecom services manager, says that when his company assists customers with upgrading a network, it performs a market-based analysis that takes the company’s existing budget and services into account. “In many cases, customers are able to purchase an Ethernet-based solution for the same cost as the existing T1 solution they have in place. This provides additional bandwidth but keeps the cost the same or even reduces it.”
Few, if any, enterprises deliberately underbuy bandwidth, but unforeseen situations can leave companies in such a position. Bernard Golden, CEO of HyperStratus, says a company might acquire another company, for example, and effectively double its traffic load. The key to planning bandwidth is not getting locked into an inflexible, long contract that disallows adding or subtracting bandwidth. Instead, Golden advises, seek out a contract that enables renegotiating terms to obtain the bulk price applicable to the added bandwidth vs. being stuck at a current price.
For many companies, obtaining the right amount of bandwidth means initially performing a detailed network assessment; something that may require a solution provider’s assistance. CDW, for example, presents customersa detailed roadmap of their bandwidth utilization and needs along with requisite hardware after performing an assessment, Shonholz says. “Understanding the applications running on the network today and in the next 18 to 24 months will also help a customer plan the best network design for its unique needs,” he says.
Failing to comprehensively assess future bandwidth requirements is arguably the biggest pitfall an enterprise can encounter, followed closely by possible hits to performance when deploying applications without fully planning how they will impact the network. After deploying applications, implementing tools to measure QoS also becomes increasingly important.