Author: Rich Hillebrecht, CIO at Riverbed Technology
Even as enterprises grow more comfortable with migrating information stores and applications to cloud and SaaS-based services, the bell does not toll for the data center. Budgetary constraints, technical limitations, security concerns and/or performance issues prevent moving the entire IT infrastructure to the cloud, and that will be the case for a while. Enterprises are therefore implementing a hybrid enterprise model, which enables them to retain control of their IT environments while sending a mix of critical and non-mission-critical workloads to the public cloud. This requires live visibility into both network and application activity. Ideally, these two commonly separate viewpoints are brought together to reveal relationships and dependencies to multiple relevant stakeholders. However, that’s easier said than done. The typical IT operations team has a compartmentalized approach to management tools and that presents significant barriers to unified visibility. Significant, but not insurmountable.
The enterprise is transforming into a mix of self-hosted applications and cloud-based services. External cloud-based applications can be a mix of SaaS applications and customer-developed applications running on external Infrastructure-as-a-Service (IaaS) platforms. With business data exchanged between on-premise apps in private data centers and other apps running in the public cloud, end to end visibility is key.
A hybrid enterprise does not reflect only where apps are hosted, but also how they’re delivered to or accessed by end users of the applications and associated data. Networks, the delivery channels that form the enterprise’s nervous system, are also going hybrid. Private networks – Multiprotocol Label Switching (MPLS) links –are being joined by public networks –the Internet – to offer choice in delivery channels: more costly private networks for mission-critical apps, less expensive public networks for bulk loads like backup. Prioritization of application traffic inside the delivery channels is also required to achieve user experience and cost-of-service objectives.
As IT decision-makers are working to implement and manage these viable hybrid networks and environments, they are operating in a business environment where application and network performance and availability are essential to generating revenue and maintaining customer relationships. Delivering applications and services across the network that make it easier for everyone to sell, buy and use products from any enterprise is a core mission for IT today. A recent Enterprise Management Associates white paper reports that IT decision makers identify “ensuring network performance” as their number one data center challenge. The primary reason for this emphasis are enterprises using service quality and application performance as key metrics of success for their network management teams to support corporate objectives.
Addressing these new business drivers is no simple task. Outages are still a common problem and can hit unexpectedly at least several times a month. Yet, as Aberdeen Group Senior Research Analyst Jim Rapoza writes in a recent report, many organizations remain in the dark when it comes to understanding their networks and the performance of the applications running on it. This makes them unable to identify and remedy issues in a timely manner, which leads to poor performance, unhappy end-users and downtime for critical systems and applications. According to Rapoza, this is why network and application performance monitoring capabilities are finding their way into other network and infrastructure hardware and systems.
“As the network becomes more dynamic and software-defined, look for multi-purpose equipment that has built-in deep and real-time visibility into network and application activity,” he adds. “This added value will provide actionable insight that companies can use to ensure that their key applications and services are running at their best possible performance.”
Barriers to Unified Visibility
A chief obstacle to realizing those improvements is the fact that network management teams use multiple monitoring tools, which adds complexity to performance management efforts. Add in the siloed nature of data center operations, and monitoring becomes even more complex.
What has led to this hodge-podge of monitoring tools? IT organizations have evolved to have multiple functional teams tasked with managing separate technology stacks, including networks, servers, and applications. Each functional team has its own set of monitoring tools to manage the technology domains inside their silos. In most cases, these tools have evolved in isolation from each other, which makes troubleshooting a tedious and time-consuming exercise, with each team using its own tools for root-cause analysis. There is no shared collaborative data set that gives an end-to-end view of services, and the source of performance problems might not be immediately obvious to a siloed management tool because the root cause might have to do with the relationships between infrastructure layers rather than with one particular part of the infrastructure.
A true unified visibility platform that pulls together network and application monitoring can solve this siloed operational model, but the path to integration is not entirely clear. These monitoring tools have evolved separately over the years, and aligning these very different data sources into something that is useful and usable is challenging.
Overcoming the barriers
A unified visibility architecture must be able to understand and present the relationships between application components and network components to provide an end-to-end view of a networked application. For this reason, IT decision makers should investigate whether an Automatic Database Diagnostic Monitor (ADDM) tool will provide a good foundation for this relationship presentation. Potential benefits include:
In today’s business environment, employees demand unfettered access to applications whether they’re in corporate headquarters, branch offices, while traveling and even at home. As importantly, business partners and customers who interact digitally with the enterprise have low tolerance for poor performance. A siloed IT approach cannot provide acceptable user experience across all the dimensions involved – and failure to deliver that experience is unacceptable in a digital age. Enterprises, therefore, need unified visibility into applications and infrastructure while taking advantage of hybrid solutions. Enterprises need network and application monitoring data presented as an end-to-end service assurance view, with insight into application dependencies and real-time data.
 Enterprise Management Associates, “The Power of Two: Why Network and Application Visibility Are Better Together,” April 2015.
 Jim Rapoza, “A Light in the Darkness: Understanding Application Performance Through End-to-End Visibility,” Aberdeen Group, May 2015.