Improve New App Implementation Using Your Old Environment
By proactively managing new technology implementations and fully understanding the old IT environment, your organization can meet end-user expectations faster while managing the overall impact to your network.
By Charles Thompson
Organizations are now moving away from traditional IT and communication architecture in favor of emerging technologies because they can increase flexibility and reduce expenses. These new technologies include data center virtualization, cloud options -- such as platform-as-a-service (PaaS), infrastructure-as-a-service (IaaS), and software-as-a-service (SaaS) -- as well as unified communications (UC) favorites voice and video.
As organizations embrace new offerings and seek new benefits, they must remember many services deployed into these environments are either critical to business or will interact with business-critical services. When implementing new capabilities, IT needs to ensure that the user experience for existing services isn't sacrificed because an established user-base brings established expectations.
When moving existing services to a new platform, how can IT teams meet expectations? We recommend using behavior analysis to understand how new and old applications perform within the context of the complete environment -- and how app and service behaviors may potentially impact users.
Use Behavior Analysis to Improve the End-User Experience
The core concepts of behavior analysis are intuitive. Organizations monitor specific activities and related variables to establish baselines, then apply these norms with alerts that notify you when anomalies occur. For IT, variables typically relate to the end-user experience (EUE). This means any condition that can impact EUE needs behavior analysis technology tracking to ensure a complete picture of the existing environment. This includes system resources, transaction and propagation delays, and response codes.
Behavior analysis offers several advantages. Here's just one example. Assume an organization is moving a key service such as order entry and inventory tracking from traditional data center/hardware architecture to PaaS. For the end user this move appears seamless, but it will fundamentally change how client-service interactions are processed. In addition to traversing the local network, clients now need Internet access to interact with the application. The likely outcome: a dramatic increase in Internet traffic, and because the Internet comes with no performance guarantees, the move can negatively impact users.
To understand the effects of change and assess existing user expectations, organizations must know what's normal for their current environment. With behavior learning, companies can predict potential issues by determining typical Internet link usage patterns and then comparing with order entry and inventory service usage patterns. This allows organizations to understand current app bandwidth demands and ensure proper sizing and prioritization for their Internet connectivity.
It's also essential to establish what's normal for overall propagation delay today to provide a contrast with end-user experience during pre-deployment testing in the PaaS environment. This guarantees strong congruency between user expectations and actual experience. This ultimately helps determine overall impact while setting proper expectations. Where applicable, leveraging local caching technologies can aid in optimizing overall EUE.
Because one benefit of cloud migration is cutting costs through infrastructure reductions, enterprises need to see into the one-on-one relationships between old servers and applications to understand capacity-use deficits. Studies show that traditional server infrastructures run at a fraction of their total compute capability. In an effort to optimize resource allocations, many IT shops will virtualize existing systems and place multiple VMs together on a single piece of hardware running a hypervisor. Compared to existing environments, this likely means a fundamental reduction in overall compute with cloudsourcing.
To ensure proper power balance and distribution, compare your baseline measurements of existing environmental behavior with those of the cloud deployment. This comparison ensures the total compute reduction doesn't result in increased processing time or create negative end-user experiences.
When deploying new UC services (including IP telephony and desktop video conferencing), latency should be consistently low to keep delivery within acceptable EUE ranges. Modern networks make it easy to segment and configure precedence for this traffic so UC has priority over more traditional services within the shared environment.
The impact of prioritizing VoIP applications ahead of other traffic is minimal due to the low amount of bandwidth it consumes. Adding high-definition and desktop video to the list of prioritized traffic changes the situation. With high-precedence communication programs taking a greater portion of network bandwidth, less space is available for critical applications such as e-mail and Web-based programs, resulting in noticeable performance slowdowns.
By applying behavior learning analysis to the pre-deployment environment, enterprises can predict performance expectations for existing services, quickly identifying any meaningful effects of UC implementation. The result? A smoother deployment for new technology and assurance that established expectations for existing services can be effectively maintained.
Visibility into Performance
Behavior learning provides organizations insight into user-base usage patterns and performance levels. By creating baselines on everything from CPU utilization to network propagation delay and application errors, companies can understand what's normal for their service delivery model. As new technologies are deployed within the environment, companies can quickly determine the impact of these changes on established levels. With alarms configured for the new baselines, teams can stay a step ahead of potential issues using comprehensive performance analysis to resolve them before they impact users.
To meet the challenges of emerging technologies, use behavior analysis to provide network and EUE clarity. By proactively managing new technology implementations and fully understanding the old IT environment, organizations can meet end-user expectations faster while managing the overall impact to their network.
Charles Thompson is the director of product strategies at Network Instruments. He is a thought leader in the areas of network and application performance monitoring with over 15 years in the industry. You can contact him at email@example.com.