You are here

Voice over IP performance

Project background :

Our client is a multi-sites health care authority located in western France. The five French sites are connected to one another via a MPLS WAN. The central resources (servers, applications, core network) are located in the HQ's datacenter. The TOP Management director, out of a desire for savings and competitiveness, has decided to switch to telephony-over-IP for all of their sites.

Client's need :

The network manager must proactively monitor inter-sites voice flows in real time, through flow's parameters and performance using RTP over IP (Real Time Protocol) such as jitter, MOS, dropped packet rate, etc. Telephony is a strategic function of the IS, and using the MPLS network as IP Carrier does cause few issues.

In particular, VoIP using MPLS requires that the telecom operator provides different Classes of Service which take the form of a certain number of "voice channels". These are billed monthly on unit use for each inter-sites connection. If the users exceed the users' traffic allowances, voice flow is treated as standard-class, flow which creates problems impacting users’ perception: poor sound quality, voice drop-out, etc.

Whenever user incident rises, the Helpdesk needs to make a quick decision about the technical situation and identify what caused it. They need alerting in order to resolve the incident quickly and restore telephony’s quality of service. Additionally, the number of allowed voice channels must be closely monitored so that the resource is optimized at all times without impacting users' telephony service.

H5 audits' answer :

An H5-300 unit is installed at the hub of the central site's WAN connections. It monitors all incoming and outgoing flows. Business indicator charts are created in order to support Help Desk in handling telephony and business application operating incidents in real time for all sites. If a call comes from a site, the time taken to display and analyze contextual graphs for that site (RTP VoIP capacity level, MOS, jitter, server response time, round-trip time, capacity, etc.) is below 5 minutes. The Helpdesk's performance greatly improves, and incidents are handled in real time with the support of H5-Performance Reporter’s intuitive, complete and simple views.

Alarms are configured to trigger when the remote sites' capacity levels, network performance, and VoIP performance reach thresholds. Alerts are relayed by e-mail, through H5-Performance Reporter and by SNMP traps, allowing network engineers and Helpdesk technicians to be informed in real time: they become proactive and master the event before it becomes too critical.

Theme by Danetsoft and Danang Probo Sayekti inspired by Maksimer