 |
Claims  |
|
|
Therefore, we claim:
1. For stations operating to exchange data through a non-synchronous data
communication network, said network characterized in that data transmitted
between said stations is subject to transmission delays of indeterminate
duration within said network, apparatus for enabling a first station in
said network to gather statistics indicative of durations of said
transmission delays in respect to data transmitted through said network
from said first station to a plurality of second stations remote from said
first station, said apparatus comprising:
means at said first station for attaching control information to data
instantly being transmitted from said first station to one of said second
stations; said control information including a data identity (ID)
associated with said data instantly being transmitted, a start time,
ST.sub.0, representing the time at which said first station began
transmitting said data instantly being transmitted, and a timing script
subject to interpretation at said one of said second stations; said timing
script requiring said one of said second stations to immediately initiate
transmission of a first return message to said first station upon receipt
of said data instantly being transmitted; said first return message
containing said data ID and said start time ST.sub.0 ;
means at said first station for saving said start time ST.sub.0 in
association with said data ID;
means at said first station for generating a first time value ST.sub.1 upon
reception of said first return message at said first station; and
means at said first station for saving said first time value in association
with said saved start time value ST.sub.0 and said saved data ID;
said saved time values being useful to develop round trip delay time
information representing the aggregate transmission delays encountered by
said data instantly being transmitted, en route between said first station
and said one of said second stations, together with the aggregate
transmission delay incurred by said first return message en route between
said one of said second stations and said first station.
2. Apparatus in accordance with claim 1 wherein:
said data instantly being transmitted is susceptible of undergoing a
predetermined user handling action at an indeterminate time after its
reception at said one of said second stations receiving said data
instantly being transmitted; and
said timing script is effective to condition said one of said second
stations to initiate transmittal of a second return message to said first
station when a said predetermined user handling action occurs; said second
return message containing said data ID and a time value ET internally
developed at said one of said second stations; said time value ET
representing the time elapsed between reception at said one of said second
stations of said data instantly being transmitted and occurrence of said
predetermined user handling action.
3. Apparatus in accordance with claim 2, wherein said first station
comprises:
means effective upon reception of said second return message for saving
said time value ET in association with said data ID and said saved time
values ST.sub.0 and ST.sub.1 ;
means effective upon reception of said second return message for generating
a second time value ST.sub.2 at the respective time of reception of said
second return message; and
means for saving said second time value ST.sub.2 in association with said
other saved time values ST.sub.0, ST.sub.1 and ET;
all of said saved time values being useful together to develop delay time
statistics indicative of the transmission delay encountered by said data
instantly being transmitted en route from said first station to said one
of said second stations.
4. Apparatus for a data server linked via a non-synchronous data
communication network to client stations, said apparatus comprising:
means for attaching control information to respective data instantly being
transmitted from said server to said network for transfer via said network
to a said client station; said control information including a data
identity (ID) associated with said data instantly being transmitted, a
start time ST.sub.0 representing the time at which said server started
transmitting said respective data, and a timing script subject to
interpretation at a client station receiving said respective data; said
timing script requiring a client station receiving said respective data to
immediately transmit a first return message to said network for transfer
to said server; said first return message containing said data ID and said
start time ST.sub.0 ;
means for generating a time value ST.sub.1 when said first return message
is received at said server; and
means for saving said value ST.sub.1 in association with the data ID and
start time value ST.sub.0 contained in said first return message; said
saved data ID and time values being useful to develop information about
aggregate delays encountered by said respective data and said first return
message in their passages through said network.
5. Apparatus according to claim 4 wherein said non-synchronous network is
one in which transmitted data is subject to routing through variably
selected switching nodes in said network and thereby subject to
indeterminate transfer delays associated with said routing and queueing
conditions encountered at individual switching nodes within a selected
route; and wherein said saved ID and time values are useful to develop
indications of said transfer delays.
6. Apparatus according to claim 4 wherein said attached timing script is
further constructed to:
a) prepare said client station receiving said data instantly being
transmitted to detect occurrence of a predetermined event which is subject
to occurring at said client station at an indeterminate time after
reception of said data instantly being transmitted; and
b) condition said client station to send a second return message of
predetermined form to said server upon detecting said predetermined event;
said second message including said data ID and a time value ET internally
developed at said client station; said time value ET representing time
elapsed at said client station between reception of said data instantly
being transmitted and detection of said predetermined event; and wherein
said data server includes means for generating a second time value
ST.sub.2, when said second return message is received at said data server,
and for saving said second time value in association with said saved data
ID and said other saved time values ET, ST.sub.0, and ST.sub.1 ;
said saved second time value and said saved other time values being useful
to develop statistically relevant indications of average transmission
delays incurred in transmitting data from said server to said client
stations.
7. A computer program product consisting of a program application for a
data server linked via a non-synchronous network to client stations, said
program application being conveyed to said data server via the computer
readable medium and enabling said data server to develop statistically
useful indications of delays incurred in transmitting data through said
network to said client stations; said program application specifically
causing said data server to:
attach control information to data instantly being transmitted from said
data server to a said client station; said attached control information
including: a data identity (ID) function uniquely associated with said
data instantly being transmitted; a start time value ST.sub.0, denoting
the time at which said server began transmitting said data instantly being
transmitted; and a timing script subject to interpretation at a client
station receiving said data instantly being transmitted; said timing
script being formed to cause said last-mentioned client station to
immediately initiate transmission to said data server of a first return
message having a predetermined form upon reception of said data instantly
being transmitted; said first return message including said data ID and
said start time value ST.sub.0 ; and
said program application is constructed to cause said first server, upon
receiving said first return message to save said start time value
ST.sub.0, in association with said data ID, as a time statistic pertinent
to determination of the transmission delay incurred by said data instantly
being transmitted en route from said data server to said client station.
8. A computer program product in accordance with claim 7 wherein:
said timing script is additionally formed to:
a) prepare said last-mentioned client station to detect occurrence of a
predetermined event; said predetermined event being subject to occurrence
at said client station at an indeterminate time after said data instantly
being transmitted is received at said client station; and
b) require said client station to initiate transmission of a second return
message to said data server upon detection of said event; said second
return message containing said data ID and a time value ET; said time ET
being internally generated by said client station and representing the
time elapsed at said client station between reception of said data
instantly being transmitted and occurrence of said event; and
said program application is constructed to cause said data server, upon
reception of said second return message, to:
a) generate a second time value ST.sub.1 representing the time of said
reception, and
b) save said second time value and said elapsed time value ET in
association with said data ID and said saved start time value ST.sub.0 ;
said saved time values together providing a basis for deriving a
statistically relevant indication of the transmission delay incurred by
said data instantly being transmitted en route from said data server to
said client station. |
|
 |
|
 |
Claims |
|
|
|
Description |
|
|
FIELD OF THE INVENTION
This invention concerns a method and apparatus for gathering timing
statistics in non-synchronous data communication networks; particularly
statistics pertaining to data transmission times and reaction times of
users at receiving stations.
BACKGROUND OF THE INVENTION
The term "non-synchronous data communication networks", as used herein, is
meant to connote networks in which internal clocks of sending and
receiving stations are essentially not synchronizable. Typical examples of
such networks are the Internet and World Wide Web; wherein transmitted
data may be subject to indefinite delays due to varied routing (depending
upon traffic conditions) and due to conditions encountered at individual
switching nodes in any route (depending upon traffic conditions and
operating speed of equipment at respective nodes). Accordingly, such
networks do not lend themselves easily to gathering of timing statistics
indicative of transmission time delays between sending and receiving
stations. Even assuming that receiving stations send notifications of
reception to origin/sending stations, such notifications would necessarily
arrive at indefinite times due to indeterminate delays caused by routing
and other conditions (like those encountered by transmitted data).
Furthermore, at receiving stations in such networks, the timing of user
reactions to received data is generally indefinite, and, assuming
receiving stations send notifications of such reaction times to
origin/sending stations, times at which such notifications arrive at
sending stations are necessarily indefinite due to network delays of the
same type as those encountered by transmitted data.
Nevertheless, statistics indicative of data transmission times over the
network and user reaction times at receiving stations may be very useful
for enabling data senders to evaluate their communication needs and the
effectiveness of their transmitted data. This is particularly true of
public networks like the Internet, in which transmitted data may have
commercial or advertising significance.
It is generally known that in such networks transmission delays are not
measurable or inferrable simply from acknowledgements returned by
receiving stations; inasmuch as transmitted data would encounter
indeterminate routing and/or switching delays, and corresponding
acknowledgements would also encounter indeterminate routing and/or
switching delays generally different from those encountered by associated
transmitted data. Similarly, it is understood that timing of user
reactions to transmitted data, at receiving stations, is not inferrable
simply from returned acknowledgements or equivalent signals inasmuch as
such reaction times per se have indefinite timing relative to instants of
reception and return transmissions of acknowledgements would be subject to
indeterminate delays due to the network construction.
Nevertheless, information about data transmission delays, and delays in
respect to user reactions to transmitted data at receiving stations, could
be quite useful. For instance, the sender of an advertisement might want
to know how long it took for the advertisement to reach the average
receiver, and what interest is generated by the received advertisement or
even specific portions thereof. Knowledge of the average transmission
delay could enable senders to evaluate their communication needs and
upgrade or downgrade accordingly. Similarly, knowledge of how times
elapsed on average between reception of advertisements and actions
indicating user interest (or even disinterest) therein could enable
senders to evaluate the effectiveness of such advertisements.
In general then, for data representing various displayable offerings,
advertisements, or the like, time statistics indicative of transmission
delays and reactions to the received data could be of interest to senders;
at least as a means for evaluating the effectiveness of transmission
facilities currently being utilized and effectiveness of the data per se.
SUMMARY OF THE INVENTION
In accordance with the invention, means located at server and client nodes
of a data communication network--particularly, a network in which data is
subject to being transmitted over varied routes depending upon time of
day, traffic, class of service, etc.--enable a computer at the server
node, when transmitting data to a computer or "intelligent" terminal at
the client node, to receive feedback messages from the client node
containing timing useful timing statistics. More specifically, timing
statistics provided in the feedback messages are useful for indicating the
types and frequencies of transmission delays that are attributable to
network routing, as well as indicating user reactions to transmitted data
at client terminals. Thus, the owner of the transmitted data can use such
statistics to determine if their network service has to be upgraded, to
select optimal times of day for transmitting their data, to determine the
effectiveness of the data per se in respect to those viewing the data at
client stations, etc.
To induce actions at client stations resulting in return of such feedback
messages, a "timing script" is incorporated into the data transmitted to
each client station. This script is interpretable by an interpreter
operating at the client station. For instance, on the Internet, the timing
script could be written in the script language associated with the
Java.TM. programming language (Java is a Trademark of Sun Microsystems)
commonly used in Internet communications, and at the client station the
timing script could be interpreted by a Java interpreter of the type
commonly provided in so-called "web browsers" like Netscape's "Navigator".
It should be understood that elements of the Java language are cited here
only as examples; and that other script languages and their interpreters
could be used for the same purpose of inducing message feedback actions.
In addition to the foregoing timing script, information added to the
transmitted data (at the server) includes a "data ID" term uniquely
associated to the data being transmitted, and a starting time value
ST.sub.0 representing the time at which transmission of the associated
data began (based on the state of the server's internal clock).
On receipt of the data at a client station, the timing script acts as
instructions to the client's interpreter program causing the latter to:
(a) immediately return a first message ("MSG1") to the server, and (b)
start a client timer ("CT") to effectively measure elapsed time from
reception of the data to one or more events specified in the timing
script. MSG1 includes ST.sub.0 and the data ID.
On receiving MSG1, the server creates a current time stamp ST.sub.1,
calculates the difference between ST.sub.1 and ST.sub.0, and saves the
result in association with the data ID as a measure of "round-trip
transmission time" taken for transfer of the data and return of MSG1. The
saved result is useful for further analysis of the effectiveness of
network access services then being received by the data sender.
Furthermore, if MSG1 is accompanied by address information indicating the
routing of that message, such address information could also be saved and
used by the sender for evaluating the transmission time taken for sending
the original data as well as the effectiveness of network services then
being used by the sender.
Upon occurrence of a said specified event, the client computer: (a)
extracts the then instantaneous value of its timer count CT as a
representation "ET" of the time elapsed between receipt of the data and
occurrence of the respective event; and (b) sends a second message, MSG2,
to the server, this message containing the value of ET as well as the data
ID and ST.sub.0. On receiving MSG2, the server creates a current time
stamp ST.sub.2, calculates the difference between ST.sub.2 and ST.sub.0,
and saves the result along with the value of ET, in association with the
data ID, for further analysis. This second saved result is indicative of
the aggregate time taken for transmission of the data, client handling of
the data up to the respective event, and return transmission of MSG2; i.e.
it is indicative of the aggregate round trip transmission times of the
data and MSG2 plus the elapsed time ET from client reception of the data
to the aforesaid event. Incidentally, it should be noted that elapsed time
between transmission of the data and multiple events, the latter
representing various reactions of a viewer to the transmitted data, could
be determined in this manner by having the client station return
additional messages MSG3, MSG4, etc.
Upon subtraction of ET from the above second result, the remainder is a
useful indication of aggregate round trip transmission times for the data
and MSG2. Also, comparison of that remainder with the "round trip" time
derived from MSG1 may provide useful insights as to differences in return
transmission routes allocated to MSG1 and MSG2. The average of these
differences in return routing can be subtracted from the remainder to
arrive at a reasonable measure of the transmission time of the originally
sent data; i.e. as a measure of the effectiveness of network service
provided for that transmission. The saved value of ET, and other values of
ET derived from other feedback messages--e.g. from messages responsive to
other events at the same client and/or from messages generated by other
clients receiving the same data--can be used to provide the data owner
(e.g. advertiser) with information indicative of various reactions to the
data of end users at client stations e.g. information that could be used
to evaluate the effectiveness of data constituting different
advertisements of a product as an aid in selection of a most effective
advertisement.
These and other features, advantages, objectives and benefits of the
present invention will be more fully understood by considering the
following detailed description and claims.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a schematic illustrating an aspect of problems dealt with
presently in respect to gathering information about network delays
encountered by transmitted data in non-synchronous data communication
networks wherein data is subject to indefinite routing and switching
delays.
FIG. 2 illustrates other aspects of problems dealt with presently in
respect to gathering information about network delays encountered by
transmitted data based on signals returned from receiving stations as soon
as the latter receive the data, and about delays in timing of user
reactions to the transmitted data, at receiving stations, based on other
sign sent when such reactions occur.
FIG. 3 is a block diagram of an embodiment of the present invention, and
its deployment at server and client nodes of a network exemplified by the
Internet.
FIG. 4 is a flow diagram describing operations performed in accordance with
the invention to effect transmission of data from a server station to a
client station, in a form enabling the server station to gather time
statistics pertaining to the time taken for the data to reach client
station as well as other time statistics presently pertinent.
FIG. 5 is a flow diagram describing operations at client stations in
respect to data received in the form indicated in FIG. 4, which operations
result in return of messages to the server station from which presently
pertinent time statistics can be calculated and saved at the client
station.
DETAILED DESCRIPTION
1. Problems Solved By The Present Invention
A network environment in which the invention can be used to advantage is
suggested in FIG. 1, and problems pertaining to gathering of delay timing
statistics in such networks are explained with reference to FIG. 1 and a
timing diagram in FIG. 2.
FIG. 1 shows a wide area non-synchronous data communication network 1
linking server 2 to client computer 3. It is understood that in addition
to the stations containing server 2 and client system 3 the network shown
may serve many other stations, both servers and clients. A presently
relevant characteristic of such networks is that data is subject to
indeterminate transmission delays due to either or both of: varied routing
of the data (through switching equipment variably located in the network)
and/or conditions encountered at network switching nodes. Such delays are
necessitated by traffic volume and other circumstances both throughout the
network and at individual switching nodes. Consequently, data transmitted
from server 2 to client station 3 would encounter indeterminate or
indefinite delays in such transmission, depending upon the number of
switching nodes traversed by the data, cumulative distances between
traversed switching nodes, conditions encountered within individual
switching nodes (traffic conditions, operating speed of switch equipment,
etc.), etc. The Internet and World Wide Web represent typical networks of
this type.
Another pertinent characteristic of this type of network is that signals
returned from stations receiving data to stations sending the data (e.g.
from station 3 to server 1 in the illustration of FIG. 1) encounter
indeterminate or indefinite network delays in their transfers that further
complicate the problem of gathering or inferring timing statistics.
Referring to the timing diagram in FIG. 2, line 4 denotes time and the
direction of increasing time is suggested by the arrow indicated at 4a.
Broken portions of line 4 represent time intervals of indefinite or
indeterminate duration.
A first time of present interest in this figure, that indicated at 5, is
the time of reception of transmitted data (e.g. in FIG. 1, data received
by client 3 from server 2). As noted in the discussion of FIG. 1, this
time is indeterminate due to network delays that depend on network routing
conditions and/or queueing delays at network switching nodes.
Now, assuming that a signal or message effectively acknowledging reception
is returned from a station receiving data to the data's origin station
(e.g. from client 3 to server 2 in FIG. 1), a second time of present
interest would be the time elapsed between the transmission of that
(acknowledging) signal or message and its arrival at its destination; i.e.
the time elapsed between time points 5 and 6 in FIG. 2. As indicated by
the broken line between times 5 and 6, this arrival time is also
indefinite or indeterminate as suggested by the broken line between time
points 5 and 6).
A third time of present interest is that elapsed between time points 5 and
7, FIG. 2. This is the elapsed time between reception of data and
occurrence of a predetermined user reaction to all or part of that data;
e.g. could be one of: discarding a viewable page of data evoked by a user
action calling for loading of new data, selection of a particular term or
symbol, in a viewable page, that effectively constitutes a "hyperlink" for
initiating a predetermined communication to the origin data server,
sending of a reply to data constituting an E-Mail note, etc.). As
indicated by the dotted line between time points 5 and 7, this elapsed
time also is of indeterminate or indefinite duration inasmuch as it
depends upon unpredictable behavior of a user/viewer of the data.
Finally, assuming that the client station sends a signal or message to the
origin server at user reaction time 7 (that signal or message effectively
constituting an acknowledgement of occurrence of the predetermined user
reaction), a fourth time of present interest would be the time elapsed
between transmission of this signal or message and its arrival at the
origin server (the latter time shown at 8, FIG. 2). As indicated by the
dotted line between times 7 and 8, this elapsed time is also of
indeterminate or indefinite duration due e.g. to unpredictable network
delays and/or unpredictable queueing (traffic) conditions at network
switching nodes.
A point to note in respect to reception times 6 and 8 is that although FIG.
2 shows respective signals/messages reaching the server in the same order
as they were sent, they could conceivably arrive in reverse order;
inasmuch as they would in general traverse different routes through
switching nodes of the network and/or different aggregate queueing delays
in transit through such nodes. Thus, the signal representing notification
of data reception could traverse a lengthier network route than the signal
representing notification of user reaction, and thereby incur a lengthier
transmission delay than the user reaction notification, and/or the former
signal could incur encounters lengthier aggregate queueing delays than the
latter signal in transit through switching nodes of respective routes.
In an environment like the Internet, representative transmitted data could
be pages displayable at the receiving client station, and a typical user
reaction that could be the subject of the above notification of reaction
could be one of: filling in a particular item in a form displayed in a
page, selection of a highlighted text term or symbol within a displayed
page (such selection e.g. effecting a hyperlink request to have a related
page downloaded from the origin server), or discarding the displayed page
by requesting another unrelated page, etc. Also, it is generally
understood that in an environment such as the Internet, sending and
receiving stations have internal clocks that are neither synchronous to
each other nor synchronizable as such; and this inability to synchronize
clocks further complicates the gathering of presently relevant delay
timing statistics.
Referring again to FIG. 2, and considering the indefiniteness or
indeterminateness of elapsed times between times 5 and 6, times 5 and 7,
and times 7 and 8, it is easy to understand why it is difficult in such
networks to gather timing statistics indicative of such elapsed times and
of transmission and user reaction delays related to such.
Nevertheless, such statistics could be quite useful. For example, a sender
of Internet pages representing commercial offerings or advertisements,
might want to know the average transmission delay encountered in
transmittal of their pages, and/or times of specific user reactions to
items contained in their pages. Knowledge of average transmission delays
could enable the sender to determine if their network access facilities
are adequate for intended purposes or need upgrading. Knowledge of user
reactions to specific page items might enable senders to modify their
offerings or advertisements to maximize user interest and ultimately
increase associated sales.
The present invention provides means for enabling senders to gather such
time statistics.
2. Preferred Embodiment of the Invention
The general attributes of the invention are described below with reference
to FIG. 3, and specific server and client operations pertinent to the
invention are described below with reference to flow diagrams contained in
FIGS. 4 and 5.
As indicated in FIG. 3, server 2', adapted in accordance with the present
invention, includes means for incorporating into (or attaching to)
transmitted data certain control information; the latter including a
unique data identity (ID) function, a transmission start time (hereafter
denoted ST.sub.0), and a "timing script" (discussed below) for inducing
equipment at the receiving station to return messages containing
statistically pertinent time information in a desired manner.
As indicated further in FIG. 3, client station 3', as presently adapted,
contains means for interpreting the above-mentioned timing scripts and for
transmitting certain return messages to servers that sent data to which
respective timing scripts were attached.
The means mentioned above preferably are embodied in application programs
installed at the server and client stations. The timing scripts mentioned
above could be written in a script language, and the interpreter for the
timing script could be contained in browser applications commonly
installed at client stations. As noted earlier, a suitable script language
for application of the invention to the Internet would be the script
language associated with the Java.TM. (Trademark of Sun Microsystems)
programming language, and a suitable corresponding interpreter facility
for client stations could be a function that is presently available in a
number of existing browser applications. For example, the Netscape
Navigator.TM. (trademark of Netscape Corporation) is one such browser
application.
Although transmission of scripted instructions is not new per se, functions
evoked by the present timing script are considered novel. It should be
noted that any script language and associated interpreter could be
employed in the present manner. It also should be noted that it would be
possible for a server to send timing scripts in any form that is likely to
be interpreted at a client station (even, for instance, in multiple
different script language forms tailored to different interpreters).
Furthermore, users of the Internet will appreciate that script, such as
that in the presently used timing script, and interpreters associated
therewith, can be installed in computers at network stations (both client
and server stations) in various forms; e.g. via computer readable media
such as storage disks or via the network itself.
Pertinent operations of server and client stations in accordance with the
invention are described next with reference to FIGS. 4 and 5.
FIG. 4 shows relevant processes performed at an Internet type server
station, during retrieval and transmission of a data page requested by a
client station. In response to reception of a page request from a client
station (action 30, FIG. 4), the server retrieves the requested page and
attaches certain information to it (actions 32, FIG. 4). The attached
information includes a timing script, an initial time stamp designated
ST.sub.0, a page identity (ID), and a first return message designated
MSG.sub.1 (or instructions for generating such). MSG.sub.1 may include the
initial time stamp ST.sub.0.
The page and attached information are sent to a network access point
(action 34, FIG. 4) at which a route (or a portion of a route) is selected
for transmitting the respective data and attachments to the requesting
client. It is understood that such a route may consist of variably located
switching nodes of the network (nodes at which data from many network
so | | |
