How to Avoid Cloud Computing Failure
How to Avoid Cloud Computing Failure
To the novice IT manager, a shift to cloud computing may appear to be the answer they have been looking for to solve many of their network problems. No longer will their team have to worry as much about large infrastructure deployments or complex server configurations. However, diving in a little deeper reveals that cloud computing can deliver a whole new set of challenges.
Through cloud computing, organizations perform tasks or use applications that harness massive third-party computing and processing power via the Internet cloud. This allows them to quickly scale services and applications to meet changing user demands and avoid purchasing network assets for infrequent, intensive computing tasks.
While providing increased IT flexibility and potentially lowering costs, cloud computing shifts IT management priorities from the network core to the WAN/Internet connection. Cloud computing extends the organization's network via the Internet, tying into other networks to access services, applications and data.
Understanding this shift, IT teams must adequately prepare the network and adjust management styles to realize the promise of cloud computing. Here are five key guidelines organizations should understand when planning, employing and managing cloud computing applications and services:
Guideline No. 1: Conduct pre-deployment and readiness assessments
Determine existing bandwidth demands per user, per department and for the organization as a whole. With the service provider's help, calculate the average bandwidth demand per user for each new service you plan to deploy. This allows the IT staff to appropriately scale the Internet connection and prioritize and shape traffic to meet the bandwidth demands of cloud applications.
Shift the Network Management Focus
Guideline No. 2: Shift the network management focus
The advantage of cloud computing lies in the fact that the burden of applications and data storage and processing can be placed on another network. This shifts management priorities from internal data concerns to external ones. Currently, organizations have larger network pipes and infrastructure at the network core where the computer processing power is located.
With cloud computing and software as a service (SAAS) applications, the importance of large bandwidth capacities shift away from the core to the Internet connection. The shift in focus will significantly impact the decisions made, ranging from whether or not an organization's monitoring tools adequately track WAN performance to the personnel and resources devoted to managing WAN-related issues.
Guideline No. 3: Determine priorities
With a massive pipeline to the Internet handling online applications and processing, data prioritization becomes critical. Having an individual IP consuming 30 percent of the organization's bandwidth becomes unworkable. Prioritize cloud and SAAS applications, and throttle traffic to make sure bandwidth is appropriately allocated.
Guideline No. 4: Consider ISP redundancy
Thoroughly assess the reliability of your existing ISP. When the Internet connection is down or degraded, business productivity will be impacted. Consider having multiple providers in case one has a performance issue.
Hold Service Providers Accountable
Guideline No. 5: Hold service providers accountable
Today, if a problem occurs within the network core, the engineer can monitor the entire path of network traffic from the client to the server in order to locate the problem source. With service providers controlling the majority of information in cloud computing, it becomes more difficult to monitor, optimize and troubleshoot connections.
As a result, service-level agreements (SLAs) take on greater importance, ensuring expected network and Internet performance levels. SLAs should outline the delivery of expected Internet service levels and performance obligations that service providers must meet, as well as define unacceptable levels of dropped frames and other performance metrics.
An SLA by itself is not enough to guarantee an organization will receive the level of service promised. Since it is not in the provider's interest to inform a client when its quality of service (QOS) fails, clients must rely on an independent view of WAN link connections. Utilize a network analyzer with a WAN probe to verify QOS and gauge whether the provider is meeting SLA obligations.
Cloud computing is more than the latest IT buzzword; it's a real way for companies to quickly obtain greater network flexibility, scalability and computing power for less money. But similar to most technologies, these services are not without potential problems. Cloud computing requires proper preparation and refocused management efforts in order to succeed.
Dave Kofflin is a Manager of Sales Engineering at Network Instruments, LLC. Dave has been building and managing networks for over 15 years. Having started out with the U.S. Marine Corps, he worked in information systems control, designing and deploying networks for the battlefield. After the military, Dave worked in various software companies, providing in-the-field application and technical support for clients. Having been with Network Instruments since 2006, Dave has in-depth experience helping network teams tackle virtually every facet of network monitoring and performance management-from application and technology rollouts to daily monitoring and problem solving. He can be reached at firstname.lastname@example.org.