When troubleshooting application performance problems, developers usually aren't on a level playing field with network managers. That’s because many traffic monitoring tools offer reports only in TCP/IP, the network manager’s lingo, according to Eric Gray, NetScout chief solutions architect. Gray helped build NetScout’s nGeniusONE Unified Performance Management (UPM) platform, which delivers application performance information in formats that networking and development teams can understand.
“Our tool gives the network team, app developer and security manager common datasets they can understand and work with,” said Gray in an interview at JavaOne 2014. “The developer gets info in the language of SQL -- not the language of TCP/IP -- from the app standpoint, not just the network view.”
What it does
The nGeniusONE UPM platform combines network and application performance management into one tool set. “Reports show every element involved in traffic flow moving across application tiers, end-to-end networks and diverse user devices,” said Jim McNeil, NetScout VP of marketing. Released in summer 2014, it was demoed at Oracle OpenWorld 2014 recently.
By delivering traffic reports that all performance management team members can interpret, nGeniusONE streamlines troubleshooting, according to Gray. Using packet-flow data, the platform mixes simple workflows with pertinent performance and quality information. Reports show hop-by-hop transaction latency, service requests and load, new and active sessions and correlated application and network errors. This service-centric approach reveals critical interactions and dependencies between the infrastructure and application tiers across complex and distributed environments, and identifies issues that host and agent-based application performance management approaches cannot.
“We built our tools to fit business-side needs, so the whole performance management team could take action and identify and triage performance problems,” Gray said.
Why it's cool for application developers
Too often, developers are fingered when application performance issues appear, because reports are delivered only in networking metrics, McNeil said. Traditionally, either one operational group tried to prove it was not their part of the network, or two or more teams pointed to each other as the source of the problem, or even third-party vendors were called out as the problem source. The nGeniusONE platform generates application and network metrics to fully expose the end-to-end performance and responsiveness of any service.
By providing one cohesive, consistent set of analytics and views, based on a unified report, developers have proof they did their job correctly, and that the problem is elsewhere. With the common database of metadata, the platform improves communication and collaboration across IT groups. “Rather than wasting time [with] the blame game, the whole team can be improving mean time to repair,” McNeil said.
What is the learning curve?
The nGeniusONE platform’s intuitive GUI design includes Web-based access for multiple users, and customizable, scheduled reports that can be emailed to other team members, according to Gray. The platform can, via a lightweight HTML5 Web browser, allow the visualization of any analysis view from any device, anywhere. Beginning with a service-focused, real-time dashboard, users can easily progress to multi-dimensional service and network traffic diagnostics, with the ability to progress into granular user-level session transaction tracing. Although the majority of issues can be resolved with higher-level metadata analysis, experts can access deep-dive packet details as desired.
How has it worked in the real world?
To provide constant and immediate access to customer and account data, the IT staff of a Fortune 500 bank had to speed up fixes to networking performance problems, according to Gray. The bank’s IT group had been troubleshooting by sending pings to infrastructure devices and interfaces. That practice ruled out a suspect location as the source of a problem, rather than showing how the issue was impacting the performance of networked business applications. The nGeniusONE’s multi-dimensional traffic reports enabled better collaboration between engineering, IT operations and architecture groups when troubleshooting network problems.
The bank has found the NetScout nGenius Probes feature particularly helpful, Gray said. In the data center, the Probes tool monitors traffic flows, performs packet captures and decodes and analyzes response time in key segments. In the bank’s 500 remote branch locations, which all access “Tier 1A” customer and revenue-affecting applications, IT incorporated the NetFlow application and conversation information supplied by the remote routers with the data monitored by nGenius Probes into nGenius' unified analysis. Using only one data source would have limited their visibility, Gray said. Combining sources offers more extensive data for deep troubleshooting analysis or for historical trending information on bank services.