FAQ

Voice quality in IP-networks may degrade at any time – this is a fundamental difference to circuit-switched telephony. Consequently, communication service providers must put tools into place to control and maintain the quality. VOIPFUTURE offers a new monitoring technology that analyzes all RTP streams in real-time. The detailed results reflect the true quality of real voice streams in the network and provide a comprehensive view of the actual service quality. This is what we call voice service intelligence. VOIPFUTURE customers benefit from this information in numerous ways, e.g. by effective troubleshooting and by being able to implement and enforce voice quality SLAs.

Does VOIPFUTURE analyze the quality of real calls?

Yes. VOIPFUTURE provides a passive monitoring solution that analyzes the RTP media streams of actual calls on all monitored links. The VOIPFUTURE system analyses all streams in real-time to produce detailed information for each actual call.

Does the VOIPFUTURE solution modify traffic in any way?

No. The probes are connected to the network through TAPs (test access ports) or monitor/SPAN ports of routers/switches. This means that the probes are non-intrusive and operate on (filtered) copies of the real network traffic.

Does the VOIPFUTURE solution inject traffic into the network??

No. The solution is fully passive. The system does not actively generate test calls or inject traffic of any kind into the monitored links.

What is the ROI?

The return on investment of our solution is typically less than one year. The solution is based on standard carrier-grade server hardware and thus the CAPEX is very low compared to traditional proprietary monitoring equipment. Our customers experience significant savings from network-wide RTP monitoring by enabling more effective operations and by reducing the number and duration of customer care calls.

Which use cases are covered by your product?

VOIPFUTURE delivers the information required to keep changing voice quality under control. This information can be applied to the following exemplary use cases:

Manage & control voice quality of IP based interconnections
Define voice quality SLAs for enterprise customers
Support customer care with advanced trouble ticket handling
Verify effectiveness of QoS mechanisms in real time
Determine effective network capacity
Troubleshoot poor network performance
Document the call quality for each and every minute

How does the VOIPFUTURE approach compare to PESQ?

Most active monitoring systems implement the standard PESQ (Perceptual Evaluation of Speech Quality) to automatically assess the speech quality of a telephony system. PESQ systems generate singular test calls and thus provide information about the end-to-end voice quality along a certain communication route at a certain time. Active monitoring systems – including PESQ-based systems – do not provide information about the service quality as experienced by real users of a telephony system in their actual calls. PESQ is standardized as ITU-T Recommendation P.862.
VOIPFUTURE’s RTP monitoring solution provides deep insight into the real network performance and user experience. For each actual call it delivers a quality metric for each stream segment and for every 5 second slice.
Sometimes there is a need to compare the quality assessments of different monitoring systems. To this end, the ITU-T Recommendation P.564 defines a way to compare the quality information of a monitoring system to a PESQ reference system. The VOIPFUTURE solution conforms to P.564 and thus its results are comparable to PESQ measurements and to other P.564 conformant products.

Do you also monitor the TDM or circuit-switched part of a network?

No. We are VoIP experts and we focus on IP-based technologies, e.g. in NGN, IMS or LTE style networks. However, our solution is able to interface with conventional monitoring solutions to obtain an integrated view of voice quality in heterogeneous networks. Furthermore, the optional packet recording feature can unveil problems in the payload that originated in the circuit-switched part of a connection.

Do you monitor the control plane/call signaling?

The VOIPFUTURE solution uses SIP signaling to correlate streams to calls. Using an advanced correlation mechanism the VOIPFUTURE solution can correlate multiple media streams to a single signaling session. Media streams include those from early media (e.g. ring tones, tariff announcements), changing codecs and those influenced by network address translation (NAT) or using the ICE/TURN technologies.

My monitoring tool also provides information about voice quality. What’s the difference?

The difference is detail and focus. Pseudo-information about voice quality is easy to obtain when evaluating RTCP (RFC 3550) reports, which is what many monitoring systems do. There are however many drawbacks to this approach, which can be summarized as follows:

RTCP reports are optional and one cannot rely on their availability, e.g. for traffic originating in other domains.
RTCP reports provide an end-to-end view and thus do not guide the process of localizing the source of a problem
RTCP reports contain little useful information unless proprietary extensions are used
RTCP reports are generated by potentially untrusted devices
RTCP reports have difficulties passing through firewalls
The frequency of RTCP reports is low, so they do not provide sufficient temporal detail
RTCP reports consume precious network resources.

In short, RTCP does not tell the whole story, if it is available at all.
VOIPFUTURE’s solution solely relies on information gathered from standard RTP streams of live customer VoIP calls. Trustworthy quality information is thus always available regardless of firewalls, customer premises and network equipment. The detailed 5 second measurements also allow applying VOIPFUTURE’s patented pattern matching technology, which helps operations focus on the real problems in a network.

Is your solution tied to specific vendors of network equipment, e.g. for SBCs, switches etc.?

No, our solution is absolutely vendor-neutral. Where appropriate VOIPFUTURE supports common vendor-specific extensions to enhance the data basis and ease integration with third party equipment.

Do you require RTCP reports to be sent by VoIP endpoints?

No. The VOIPFUTURE solution measures actual RTP streams and does not rely on RTCP reports. The solution is vendor- and configuration-agnostic.

What does “full line rate” performance mean?

By full line rate performance, we mean that we analyze every single RTP packet on a link. This is different from other monitoring solutions, which analyze only a part of the packets.

Which codecs do you support?

The VOIPFUTURE system does not analyze the payload of RTP packets. Analysis of the network performance is completely independent of the employed codec. Special codec support is only required to assess the user experience by computing mean opinion scores (MOS). Currently we process G.711, G.722, G.723.1, G.726, G.728, G.729 and GSM and dynamic payload types. AMR and AMR-WB will be supported from Q4/2011.

What is the bandwidth requirement on the uplink from the probes to the central Application Manager?

The VOIPFUTURE probes analyze RTP streams and create quality data records (QDRs), for every 5 second segment. These QDRs are compressed to minimize the bandwidth requirements on the uplink to the central Application Manager. The actual compression rate depends on a number of factors, such as the average call duration and the actual quality of the voice service. From our experience, the uplink data can typically be compressed to 0.05% of the monitored link bandwidth. So, a fully utilized 1 GbE full-duplex link would generate roughly 1 Mbps worth of traffic.

This compares favorably to the overhead induced by RTCP messages, which consumes up to 100 times more network resources than VOIPFUTURE's solution. Apart from this, RTCP messages consume valuable bandwidth in the productive voice network. In contrast, the communication between VOIPFUTURE Probes and Application Manager typically utilizes a dedicated management network.

Does your solution cope with SRTP, i.e. encryption of the media plane?

Yes. SRTP only encrypts the RTP payload and VOIPFUTURE’s RTP stream analysis does not depend on the payload of an RTP packet. Adaptive codecs, such as AMR, are an exception to this rule, as information on the current mode of the codec is stored in the codec header. For such codecs MOS values cannot be calculated if encrypted.

Is your solution able to record RTP streams for offline analysis?

Yes, this is an optional module. The RTP packets are stored in the common pcap format, readily available for post-processing with common troubleshooting tools and data basis for customer care tickets with hardware vendors.

What about privacy and data protection when monitoring calls with your solution?

We do not analyze the spoken voice (payload) which ensures the protection of privacy. VOIPFUTURE stores only statistical data necessary for a detailed analysis of your voice service. Every data transfer in between Probes and the central Application Manager is secured with standardized encryption protocols.

How are MOS values calculated?

MOS calculation is based on the E-Model defined in ITU-T Recommendation G.107. MOS and R-Factor values are calculated for every 5 second segment of an RTP stream, using a sliding window. MOS and R-Factor are not available for the first segment as the E-Model prescribes at least 8 seconds worth of speech data.

What is a TAP?

Test access ports (TAP) are used to provide passive network and security monitoring applications with access to a network link. A TAP can be set up between any two network devices, such as firewalls, routers, and switches. For RTP monitoring TAPs are typically deployed close to SBCs, MGWs, IBCFs and core routers. Conceptually, TAPs are passive devices that do no influence monitored links.
A cost-effective alternative to TAPs are SPAN or Mirroring Ports of network switches. These send a copy of all packets seen on a switch to a dedicated switch port. VOIPFUTURE recommends the use of TAPs as Mirroring Ports do not always behave as expected, particularly when there is high load on the network.

Do you have a preference for a TAP vendor or model?

No.

Do you support unattended operation?

Definitely.