The article discusses:
- Changing network requirements in the digital age
- Shortcomings with the current network performance management approaches
- The complex NPM vendor landscape
- A holistic NPM architecture with an app-centric approach
- The winning aspects of this innovative NPM architecture
Regardless of the industry, organizations embarking on a digital journey are moving away from the traditional IP address/packet-centric approach to an application-centric approach to enhance productivity and customer experience. This puts the onus on the network to deliver the best experience to the end users/customers accessing these applications. To achieve this objective, network engineers need access to tools and solutions that provide in-depth network performance visibility, with a focus on application experience.
Network architecture in the LAN, WAN, and datacenter is also undergoing transformation with the growing popularity of software-defined technologies like Software Defined Access for LAN, Software Defined WAN, and Software Defined Networks for the datacenter. Here too, the focus is on applications rather than the IP packet. Similarly, enterprise IT infrastructure is experiencing unprecedented changes. Virtualization, hybrid cloud computing, mobility, collaboration and infrastructure consolidation, are driving increased adoption of application-centric approaches such as Software as a Service (SaaS) and infrastructure as a Service (IaaS). Clearly, there’s a growing need for unified view around application visibility and monitoring, service assurance, real-time troubleshooting and analytics - to enable enterprises to fulfil their digital customer experience promise.
In Feb 2019, Gartner estimated the size of the NPMD tool market to be $2.2 billion, growing at a compound annual growth rate of 1.8%1. This market includes application performance monitoring (APM), IT infrastructure monitoring, artificial intelligence for IT operations platforms, and digital experience monitoring.
There’s a growing need for a unified view around application visibility and monitoring, service assurance, real-time troubleshooting and analytics to enable enterprises to fulfil their digital customer experience promise.
How current network performance management approaches fall short
Ensuring superior application experience requires real time visibility and performance analysis into how an application is handled in the network and the possible reasons for performance degradation. Traditionally, organizations have used various tools and network utilities like ICMP, SNMP and Flow mechanisms. However, these tools offer a basic view and are not always capable of providing the necessary information for network performance assurance. They also tend to operate in silos, adding to operational complexity and increasing the time to troubleshoot issues. While information like port status, device utilization and link status are important, they don’t enable application visibility in the data center – a critical factor for measuring application performance and providing exceptional user experience. The traditional reactive approach to application performance is not sufficient for organizations embarking on a digital transformation journey. They need a granular view into their network, with an application-centric model that offers a unified view.
At the same time, Cloud adoption is rapidly increasing, with organizations moving from private only cloud to public and hybrid clouds. This, in turn, is driving new ways of datacenter networking, WAN connectivity, and access networking in enterprises. Traditional monitoring tools fall short as they are not well equipped to handle the shift in data patterns in the complex private/public cloud setup, leading to potential security vulnerabilities. While multiple vendors offer tools/solutions for visibility and analytics, organizations need a holistic approach that eliminates vendor lock-in with proprietary hardware/software.
While multiple vendors offer tools/solutions for visibility and analytics, organizations need a holistic approach that eliminates vendor lock-in with proprietary hardware/software.
Issues with the solution landscape
In the current market, there are multiple protocols for gathering all the traffic as-is or in the form of flows (like SPAN/ERSPAN, Netflow, J-Flow or S-Flow and SNMP). There are also several vendors offering varied features and functionalities. Some vendors offer “taps” - either physical or virtual - that tap into the traffic passing through the physical/virtual infrastructure, and send the flow data to tap aggregators or packet brokers. These aggregators in-turn, send the aggregated data to a centralized tool for analysis. Some vendors require network components to send Netflow, J-Flow or S-Flow - either to a packet broker or directly to the upstream tool/analyzer. Others might require configuring SPAN ports at strategic points in the network, with all data sent to them.
Adding to the complexity is the fact that the underlying infrastructure in an enterprise is not always provided by a single vendor. Multiple infrastructure components have their own ways of sending application flows to the upstream tool or analyzer. Managing multiple infrastructure components, protocols, tools and analyzers is time consuming and complex. It defeats the purpose of sending application flows to an analyzer to make troubleshooting easier and faster.
The need for innovation in NPM
An application-centric network requires next-gen tools that understand the application landscape through tight integration with the underlying infrastructure, providing the ability to understand various flows prevalent in the network. Deploying an innovative network performance management (NPM) architecture that uses existing tools, wherever possible, and brings in new tools and solutions to close the gaps in visibility and analysis (See Figure 1) is the ideal way to achieve this.
Using such architecture, organizations can retain their traditional methods of gathering application flow data like Netflow, SPAN and physical/virtual taps, and centralized tools/analyzers. All they need to do is deploy an orchestration/intent layer. The orchestration/intent layer manages all the underlying tools provided by multiple vendors and creates a single pane of glass for ease of management and analysis. They can choose a custom orchestration layer based on their existing environment and business objectives.