Multiplex communications – Communication techniques for information carried in plural... – Combining or distributing information via time channels
Reexamination Certificate
1999-10-04
2003-06-10
Vincent, David (Department: 2661)
Multiplex communications
Communication techniques for information carried in plural...
Combining or distributing information via time channels
C370S352000, C370S353000, C370S354000, C370S395200, C370S395210, C370S395420, C370S469000, C370S473000, C370S509000, C370S400000
Reexamination Certificate
active
06577648
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates to a method and apparatus for determining Voice Over Internet Protocol (VoIP) Quality of Service (QoS) characteristics of a network using multiple streams of packets, wherein measurements obtained as a result of the multiple streams of packets are synchronized relative to each other to permit the QoS characteristics of parts of a network to be reliably determined.
Voice or telephony services can now be provided over a packet switched network, such as the Internet. These packet switched networks are commonly referred to as Internet Protocol (IP) networks. Delivery of telephony in IP-based networks is called voice over IP (VoIP), because the Internet Protocol according to various IP based standards is the primary bearer protocol used. One such IP based standard, for example, is the International Telecommunication Union (ITU) H.323 Standard which provides a foundation for audio, video and data communications across IP networks.
VoIP is a cost effective means of transporting digitized audio signals. However, in order to provide VoIP minimum guarantees must be met where packets containing the audio signals will arrive within a set delivery time and will not be discarded due to queue overflows. For example, some audio and video “play-back” applications are intolerant of any packets arriving after play-back time. Further, some applications have hard realtime requirements in order to operate properly. Thus, the ability to provide at least some kind of guaranteed services is needed. These guaranteed services are provided in packet switched networks in the form of Quality of Service (QoS).
The original IP standard provides no QoS support in the form of limits in packet delivery delay or lost packets. Accordingly, variations in end-to-end packet queuing delays and lost or discarded packets can occur due to changing network conditions. To remedy this, QoS standards have been proposed such as Differentiated Services (DiffServ) standards disclosed in “Definition of the Differentiated Services Field (DS Field) in the IPv4 and IPv6 Headers” by K. Nichols et al, RFC 2474, Network Working Group, IETF, December 1998 and “An Architecture for Differentiated Services” by S. Blake et al, RFC 2475, Network Working Group, IETF, December 1998. Also QoS standards have been proposed such as Integrated Services (IntServ) Standards as disclosed in “General Characterization Parameters for Integrated Service Network Elements” by S. Shenker et al, RFC 2215, Network Working Group, IETF, September 1997. When the QoS for a data stream is guaranteed, limits can be set on end-to-end delays and packet loss. Further, the over provisioning of network resources in small scale applications can be avoided.
Whatever method of providing guaranteed QoS for VoIP is used, measurements are an invaluable tool in determining whether the QoS guarantees are being met. This is especially true for IP networks in which data and voice streams co-exist.
Various methods have been proposed for measuring QoS characteristics in packet switched networks. Such measurements have been performed to address issues like router stability, distribution of end-to-end delays and lost packet rates. These proposed methods are primarily concerned with measuring QoS characteristics on a single stream of packets such as that illustrated in
FIG. 1
, wherein a sending host
101
and a pinging host
102
are provided.
As shown in
FIG. 1
the sending host
101
sends a (single) stream of packets across the packet switched network
103
. Prior to sending, the sending host
101
adds information to each of the packets. The stream of packets sent by the sending host
101
is received by the pinging host
102
. The pinging host
102
adds information to each of the packets and returns the packets to the sending host
101
across the packet switched network
103
. Thus, the system illustrated in
FIG. 1
allows for round trip measurements of QoS characteristics to be performed. The round trip measurements have the advantage of avoiding complications arising from the synchronization of the of the clocks applied to the sending host
101
and the pinging host
102
, and providing an efficient emulation of real full-duplex VoIP conversations.
An example of the format of a packet used in a round trip measurement of QoS characteristics is illustrated in FIG.
2
. The packet
200
as illustrated in
FIG. 2
has various fields for storing information added to the packets by each of the hosts. For example, fields
201
and
202
store time stamp data TS
1
and packet sequence number data SEQ
1
respectively. The TS
1
, and SEQ
1
data are entered by the sending host
101
prior to sending the packet
200
. Fields
203
and
204
store time stamp data TS
2
and packet sequence number data SEQ
2
respectively. The TS
2
and SEQ
2
data are entered by the pinging host
102
when returning the packet to the sending host
101
. The pinging host
102
returns the packet
200
to the sending host
101
. Field
205
stores time stamp data TS
3
which is entered by the sending host
101
upon receipt of the packet
200
returned by the pinging host
102
to indicate the time the packet
200
completed its round trip.
By use of the above described information stored in the fields of the packet
200
having completed the round trip, QoS characteristics of the portion of the network upon which the packet completed the round trip can be determined. QoS characteristics of the portion of the packet switched network are at acceptable levels when limits such as end-to-end delay and packet loss are not exceeded.
A high end-to-end delay lowers the QoS characteristics of the stream of packets as experienced by an end user. A “high quality” QoS may set end-to-end delay to be less than 250 ms, whereas “best” QoS may set end-to-end delay at 150 ms. Delays exceeding a target value might be interpreted as losses, resulting in “limited delay”. Jitter, or variation of packet inter-arrival times at-the receiving host is important for receiver play-back buffer dimensioning. Thus, determining a value of jitter is important. The effects of packet loss on QoS is more complexed. A tolerable packet loss rate depends on the coder/decoder (CODEC) used since such a CODEC may employ QoS counter-measures such as Forward Error Correction (FEC). An important factor affectin QoS of a packet stream is the correlation of losses which are measured as conditional loss probability distribution clp(&eegr;) (i.e., the set of probabilities P(&eegr;) that n adjacent packets are lost) and scalar conditional lost probability clp (i.e., the probability that the loss of a packet is followed by another).
Thus, the measurement of QoS characteristics determined based on the information stored in the fields of packets that have completed the round trip may include average delay-limit Round Trip Time (RTT) delay (dr
l
), delay-limited loss (l
l
), jitter (j) and average length of sequences of lost packets (plg). The term “delay-limited” loss l
l
refers to packets exceeding a certain limit, for example, 250 ms. Such packets are counted as losses and ignored in computing dr
1
.
The above described method of calculating QoS characteristics using a single stream of packets suffers from various disadvantages. For example, performing measurements on a single stream of packets in a network is not a reliable measure of the QoS experienced by various users in various parts of the network. Since only a single stream of packets is used for measuring QoS characteristics, the measurements are only relevant for the portion of the network upon which the stream of packets perform the round trip. Further, the QoS characteristics may be affected in different ways in other parts of the network.
Further, co-located measurements when the sending hosts are located in one part of the network and pinging hosts are located in another part of the network yield more reliable information. In such an arrangement, QoS characteristics of different parts of the entire network or in different parts of
Räisänen Vilho
Rosti Jari
Antonelli Terry Stout & Kraus LLP
Nguyen Van
Nokia Corporation
Vincent David
LandOfFree
Method and apparatus for determining VoIP QoS... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for determining VoIP QoS..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for determining VoIP QoS... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3118838