shortcomings of UDP,
RTP
(Real-time Transport Protocol) make the transmission of packets (this time voice or video packets) as reliable as TCP.Finally, on top of RTP we have all the VoIP protocols. Examples are
SIP
(SessionInitiation Protocol),
H323, SCCP, MGCP.
These protocols are responsible for establishing, controlling, administering and ending phone calls. Aside from these, wehave
Codecs
that represent the voice or video format. The quality of the audio streamand the bandwidth consumption depends on what and how codecs are used
. G.711
,
G.722
,
G.726
,
G.729
are
G.722
are the most used.
3.Designing and implementing a VoIP network
When designing a VoIP solution there are some things that you need to take intoconsideration. Current and needed infrastructure, number of users, budget, need featuresand even legal issues.First of all you need an
IP infrastructure
. This means that you have to have anetwork that can route IP packets. Most of this infrastructure represents equipment:
switches
(or hubs) for LAN connectivity,
routers
for WAN connectivity,
access points
for wireless capabilities and
cables
to connect these components. Also you also need toconsider things like
firewalls
(for security)
NAT servers
and DHCP servers
(for addressing management) and
user management servers
(such as LDAP or ActiveDirectory). It is important to have a “VoIP ready” infrastructure because this has animportant impact on
QoS
(quality of service) of the voice network QoS in a voice network means that you have to take care of some factors
•
Bandwidth –
the speed between devices
•
Delay –
the time a packet needs to reach the destination
•
Jitter
– the time between two packets
•
Packet loss –
number of packets that are lost in the transmissionThe bandwidth is proportional with the number of users and the codec used. If you have a large number of users you need a high bandwidth or use a codec that uses less bandwidth (but the quality of the sound will decrease).Delay is important because voice packets need to arrive in real-time. You don’twant to have moments of silence on your phone until you get a response from the other side. Rules say that you need to have a delay of maximum
150ms
.Because of the buffer delay in equipments, the packets arrive with a delay to eachother. This is called jitter. You don’t want jitter because, like in the case of delay, youhave real-time transmission and you don want words to arrive at the right time ( with nosilence time or overlapping streams).The worse case is that you have packets dropped on the way to the destination.This would mean that whole words would be lost from the conversation. This isunacceptable so the rule is that the maximum number of packets lost is 2% (2 for every100 packets sent). This is because a very small number of lost packets can be correctedusing some advanced predictions and correction methods.
Add a Comment
joel889left a comment