Multiplex communications – Communication techniques for information carried in plural... – Adaptive
Reexamination Certificate
1998-10-16
2004-02-24
Ton, Dang (Department: 2666)
Multiplex communications
Communication techniques for information carried in plural...
Adaptive
C370S328000, C370S329000, C370S330000, C370S341000, C370S349000
Reexamination Certificate
active
06697378
ABSTRACT:
CROSS REFERENCES
For background information, this patent refers to the following publications,
[Ref-
1
] “TCP and Explicit Congestion Notification”, Sally Floyd, Lawrence Berkeley Laboratory, Berkeley, Calif. 94704.
[Ref-
2
] “On the Capacity of a Cellular CDMA System”, Gilhousen, et. al., IEEE Transactions on Vehicular Technology, Volume 40, No. May 2, 1991.
[Ref-
3
] “Erlang Capacity of a Power Controlled CDMA System”, Viterbi, et. al.
BACKGROUND
1. Field of Invention
The invention relates to control of packet transmissions of various classes of IP/data traffic and obtaining/using capacity estimates from a wireless network. In particular the invention relates to,
1. estimating the available link bandwidth in a wireless network using messages (standard and proprietary) from the wireless network and/or other estimates from an external device, and
2. combining the estimate with priority/class-based control of data packet transmissions to provide an efficient use of available bandwidth (capacity) per cell (sector) on the air interface (wireless link).
2. Description of Prior Art
In order to describe the background of this invention, details on the following two areas need to be presented:
1. Bandwidth management and transmission control in data networks such as IP networks, and
2. Real-time capacity and loading in a wireless network, and the overall network architecture
TCP/IP (Transport Control Protocol/Internet Prototol) is a very well established, widely used standard protocol for data/voice communications over packet networks [REF-
1
]. TCP handles each connection independently and maintains an end-to-end flow control. As shown in
FIG. 1
, the sender (
103
) and the receiver (
103
) maintain an end-to-end TCP (
104
) peer relationship. Whereas both the sender and the receiver maintain a IP layer (
105
) relationship with the router (
101
). The TCP employs retransmissions and window size control on each data connection making it reliable even when the underlying routing/switching systems experience congestion or temporary failures. The physical layer (
100
) and layer
2
(
106
) in such networks can be various standard and/or proprietary methods such as, ethernet, frame relay, ATM, SONET, T
1
, etc. Though the TCP and IP protocol layers are independent of the lower layers, their performance and efficiency depends on the lower layers.
Several methods of implementing and improving TCP flow control have been developed, but the most basic method is the window-based flow control. When TCP on the receiving machine sends an acknowledgement, it includes a window update in the segment to tell the sender how much buffer space the receiver has available for additional data. The window update specifies the amount of data the receiver can accept beyond the data being acknowledged. The TCP sender sends the amount of data indicated by the window size. The TCP sender also estimates the round trip delay which is used to set/control the TCP window size, which implicitly controls the TCP rate. A window update of zero completely halts the sender transmission. Transmission is resumed upon receiving an acknowledgement with a non-zero window size. In general, the TCP protocol uses the window updates along with other algorithms to control the flow and avoid congestion across the connection.
Current TCP/IP networks rely on packet drops as an indication of congestion. Upon experiencing packet losses, the TCP sender retransmits the lost packet and lowers its window size to reduce the amount of data being sent at a time.
An indirect method to control the data flow of TCP connections is to introduce controllable queues in the transmission path. One such example would be an IP queue in an IP router (
101
). By queuing (delaying) IP packets, the measured roundtrip delay increases (with potential for TCP timeout), which automatically reduces the TCP window update, thereby lowering the effective TCP data rate. Such methods can be applied to various classes of IP traffic with pre-defined rules. New enhancements include methods to use Explicit Congestion Notification (ECN) [REF-
1
]. ECN is done by sending a one-bit notification to the sender indicating congestion. The sender TCP then reacts to the ECN bit by lowering the TCP window size to one and initiating a slow-start session.
The TCP window and other capabilities of the TCP/IP protocol provide the necessary capabilities to implement various bandwidth management algorithms.
The other aspect of the background is the wireless network architecture, and the associated capacity and loading.
FIG. 2
shows a typical cellular/PCS wireless network and its key components. A typical cellular network (
200
) covers a contiguous area that is generally broken down by a series of cells (
201
). Each cell has a base station (
202
) and may be subdivided into sectors. The base-station maintains a radio link with the mobile station (
203
) (eg. a cellphone, or a fixed wireless terminal, or a handheld wireless computing device). The other system elements include a Mobile Switching Center (MSC) (
205
), Base Station Controller (BSC) (
204
), and a data InterWorking Function (IWF) (
206
). The data IWF (
206
) is the entity that provides connectivity of the wireless network (and mobile stations) to the IP/data network via circuit switched and packet switched wireless data protocols.
The cellular network layout provides coverage and serves the mobile and fixed wireless stations with a wireless link to the cells (sectors). The wireless, RF (radio frequency) link to the cells could be based on established industry standards such as IS-
54
(TDMA—Time Division Multiple Access), IS-
95
(CDMA —Code Division Multiple Access), and GSM (Global System for Mobile Communications), or new upcoming standards such as cdma
2000
and WCDMA, or proprietary radio interfaces. Typically a cell (sector) is able to support a certain number of wireless calls. This capacity, number of simultaneous active calls per cell (sector) is a function of (depends on) various factors such as frequency reuse, carrier to interference ratio, bit-energy to noise ratio, effective bit-rate per call (voice or data), frame error rate (FER), etc. Several studies have been done in estimating the air link capacity in a wireless network [Ref-
2
, Ref-
3
]. The radio spectrum (frequency band) used in a particular cell (sector) is reused in every “n ” cells (sectors). For example, in a CDMA (IS-
95
-based) system, n=1 indicating that the frequency band is being re-used in every cell (sector). In other cellular systems such as GSM and TDMA, “n ” could be 3, 4 or 7 or any fraction thereof. The frequency reuse factor “n”, carrier to interference ratio (C/I), bit energy to noise ratio (Eb/No), processing gain, handoff gain, total time-slots, total frequency channels, total power, expected data rate per user, expected power per user, and engineering specifications (amongst other factors) determine the maximum capacity per cell (sector) that can be supported to ensure service within certain performance metrics. In some systems, such as CDMA (IS-
95
, W-CDMA, cdma
2000
), the capacity per cell (sector) indicates a threshold such that the system or call performance degrades below a certain quality of service if the traffic in the cell (sector) exceeds the threshold. Typically cellular/PCS wireless networks are engineered such that the number of simultaneous active calls per cell (sector) is maintained below a certain threshold. This is done to ensure acceptable system and call performance.
Radio spectrum being a limited resource, considerable engineering and technology effort is spent to ensure the most efficient use of the air interface. Fundamentally the air interface capacity per cell (sector) is limited by the data rate (in bits per second) that can be transferred across the air link for a given set of quality (FER) and reuse parameters. Also, well established traffic engineering/planning approaches are used to engineer the cellular networks such that the nu
Cisco Technology Inc.
Ton Dang
LandOfFree
Method and apparatus for class based transmission control of... 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 class based transmission control of..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for class based transmission control of... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3339771