Qosmos NFV Probe
As operators transition to NFV, monitoring solutions need to adapt to this new environment. The traditional approach using taps, splitters and physical probes does not provide the required visibility into a virtualized infrastructure. Physical probes cannot access the logical interfaces between internal VM-to-VM communication to monitor functions hosted on the same server, or in some cases between VMs located on different servers connected via virtual overlay networks. Moreover, new virtualization techniques, such as containers that rely heavily on distributing traffic between many lightweight, even transient, virtualized resources only exacerbate the visibility problem.
The Qosmos NFV Probe has been specifically developed to provide visibility into virtualized infrastructure. At the heart of this technology is Qosmos ixEngine, the market leading deep packet inspection engine providing detailed real-time IP traffic classification and metadata extraction. As shown in figure 1 below, the probe can be integrated in two ways:
- Software instance on a hypervisor monitoring the entire NFVI layer (resources and network)
- VNF running on a virtual machine to monitor specific VNFs
Figure 1: Implementation example based on NFV reference architecture
Qosmos NFV Probe can be deployed to monitor both the external traffic on the virtualization platform’s physical interfaces, as well as VM-to-VM communications. As shown in Figure 2, it offers standard APIs to support real-time online and offline views to higher-level analytics and management functions. Monitoring can be performed continuously or on-demand, depending on needs and data storage capabilities.
Figure 2: Interfaces to other systems
High-degree of information granularity for advanced troubleshooting of faults and performance issues
Fully customizable extraction of traffic information and KPIs
Optimized and dynamic configuration: monitoring can be turned on and off to investigate issues only when needed
A single monitoring approach for hybrid environments where network services run across both physical and virtual functions
Data model optimized for correlating network counters (response time, QoE, etc.) with infrastructure elements (VMs, containers, domains)
Provides a view into shadow IT applications and traffic
Follows NFV standards for future proofing
Runs on commodity hardware