Re: Help - Timing Logic

Ok. Just in case you're interested, here's how we did something that sounds
very very similar to what you are trying to achieve.

Our requirement was to have a central message lodgement resource that had to
be able to handle up to 10,000 messages during a 9 hour working day. A given
message would be for 1 and only one user but we had to cater for up to 200
users.

Messages were automatically 'lodged' by an application that could be run by
any user in the system.

Any given user may not logged in at any given time, but any given user could
be logged into multiple computers simultaneously. This meant that to deliver
a message to a given user we had to to deliver it to all machines they were
logged into. If a user was not logged, we, of course, could not deliver a
message to them but it had to be delivered to them at the first avaible
opportunity.

The way we approached this was to have a client application that ran in the
users start menu. The application, when it started, 'registered' the user to
the message lodgement resource along with the machine name. When the
application terminated, (i.e. when the user logged off), the application
'de-registered' the user/machine combination.

The user also lodged messages for other users via this application
(including users who were not online).

The application also set up a TCP Listener on a predefined port for
receiving messages.

The message lodgement resource, comprised a SQL Server 2005 database and a
server application, both of which ran on the same box.

One part of the server application was dedicated to receiving messages from
the client applications and 'lodging' them in the database.

Another part of the server application was dedicated to retrieving messages
from the databse and distributing them to the clients.

A third part of the server application was dedicated to displaying
throughput and status information.

The database comprised 3 main tables. One table held static user
information, a second table held dynamic user/machine information and the
third table was the message store.

The structure of the message store table was, essentially:

create table messagestore(
ms_id bigint not null identity(1,1),
ms_timestamp datetime not null,
ms_sender nvarchar(50) not null,
ms_recipient nvarchar(50) not null,
ms_message nvarchar(max) not null,
ms_status int not null constraint df_messagestore_01 default (0),
constraint pk_messagestore primary key (ms_timestamp,ms_id)
)

The ms_id column was used purely to keep messages in the sequence that they
were lodged if the value for the ms_timestamp column was not unique.

When a message was lodged, a value for ms_status was NOT specified in the
insert statement therefore it started 'life' with a status of 0 whaich meant
that it was available for sending.

The retrieval of messages was done via a stored procedure as follows:

create procedure retrievemessage as

set nocount on

declare @id bigint

begin transaction

select top 1 @id=ms_id from messagestore where status=0 order by
ms_timestamp,ms_id

update messagestore set ms_status=1 where ms_id=@id

select
ms_id,,
ms_timestamp,
ms_sender,
ms_recipient,
ms_message
from
messagestore
where
ms_id=@id

commit transaction

go

The server transaction had a thread with a loop that continually exectued
that stored procedure via a datareader. If there was no row returned then
the thread slept for 50 milliseconds before continuing the loop. If a row
was returned, the thread dealt with the row and continued the loop
immediately. If there were 10 iterations of the loop without there being a
50 ms sleep then a 50 ms sleep was forced. This allowed the thread to 'do
its stuff' on a timely basis with out becoming unresponsive to any control
'commands'.

The prime intent was to send all available messages in the message store as
quickly as possible after they were lodged.

The thread spawned very short-lived worker threads to do the actual sending
passing the necessary values as parameters.

A single instance of a worker thread handled one and only one message.

The worker thread determined if the user was logged in. If so it sent a copy
of the message to every machine that the user was logged in to. If it
sucessfully sent the message it updated the ms_status value in the
appropriate database row from 1 to 2. If it was unable to send the message
for any reason it updated the ms_status value in the appropriate database
row from 1 to 3.

When we stress tested this we had various users logged into a toal of 20
machines and we preloaded the message store with 50,000 messages, (5 days
worth), for those user but about 5,000 messages were for users who were not
logged in. Two of the users were logged into 5 machines each. We then
started the retrieval thread.

The time it took that thread to deal with all 50,000 messages was just over
24 seconds. Of all the worker thread instances, the longest lived thread was
just over 300 milliseconds, and, as expected, that was one ofthe threads
handle a message for one of the users that was logged in to 5 machines. The
average length of the ms_message value was about 200 Kilobytes.

Certainly the hardware we were dealing with was reasonably beasty but we
were more than happy with the performance of the techniques we used, (we had
not expected it to be quite that good).

So the moral of the story is: Don't get all bitter and twisted about worst
case performance expectations. Give it a suck and see and you may find that
your fears about performance issues might be groundless.

Good luck!!!!!


I have a multi threaded VB.NET application (4 threads) that I use
to
send text messages to many, many employees via system.timer at a 5
second interval. Basically, I look in a SQL table (queue) to
determine who needs to receive the text message then send the
message to the address. Only problem is, the employee may receive
up
to 4 of the same messages because each thread gets the recors then
sends the message. I need somehow to prevent this... just can't
think of how. Somehow I need the other threads to know that
another
thread is already using that record and move on to the next
record.
I thought of getting the record then marking it (column value) as
in
use and writing the code to look for records only where not in
use... problem is all 4 run fast enough to all use/mark the row.
Any thoughts?

"Jay" <someone@xxxxxxxxxxxxx> wrote in message
news:OGJfd9xEHHA.3188@xxxxxxxxxxxxxxxxxxxxxxx
Thanks Jeff, answers under your questions...


"jeff" <jhersey at allnorth dottt com> wrote in message
news:uekJE1xEHHA.2356@xxxxxxxxxxxxxxxxxxxxxxx

I know it is a time critical system ... employees need the information
ASAP... but what I do not understand is ...

- where is the application running ... on a server ... on a users work
station?

Server (2003)

- who will be responsible for executing / sending messages ... one
computer or many different computers?

One computer. One multithreaded vb.net app.

- will this message sender reside on a server / database server / or a
user's workstation ... ???

Remote SQL Server 2005 (EE)

- what database are you using ???

SQL Server 2005 EE

- is this functionality wrapped up in a large application?

Not very large. The logic around this loop cycle is pretty much the
entire application.

- why 4 threads ?

Allows sending 4 messages at a time.

- how does you program determine which message to send ... parameter list
please.

Read from a table in the db server. A Queue table. (employeeid,
messageaddress)




------------------------------------
If this information is so mission critical, and employees must get their
message ASAP ... put a trigger on the database table.

Cant have a trigger execute a function in a remote vb.net app.


... wrap the necessary functionality is a small stand alone exe
application

Exactly what I need to do.


...
... build your application so it receives commandline parameters / string

Considering it.

...
... build a trigger on the database .. fire on inserts ... have it call
you message program with a command parameter (messageID) ...
... install you application on the database server .... this will speed
up the connection / retrievals and so on ... no network latency.

Each time a message is inserted in the table, the trigger fires, calling
your EXE with the appropriate parameter string, exe starts, fires off
message ... done. Small exe can run as many times on the server...

------------------------------------

If you are using a timestamp for determining which messages to send ...
incorporate another table ... LastTimeExecuted ...

table: MessengerExecution
field: LastDateTimeFired

begin transactions ... lock table.
select lastdatetimefired from this table...in a variable
update field with thread date/time ...
end transaction.
return the select value and the update value

select messages where date is between lastdatetimefired and the
updateddatetime I just used...

------------------------------------

Again, there are many solutions ... just do not completely understand
what you are doing...

------------------------------------
Bottom line ...

- you have determined that you need to run a mutli-threaded process for
this.

- so, in order to avoid DUPLICATE MESSAGES .... you either have to
employee...

Transaction and Database Locking - look at isolation levels / settings

or

A booker type of system...

------------------------------------
have the broker continually pool the table ... this will only work it 1
machine is designated for messaging! If more than one machine will be
used for messaging ... you will have to roll up your sleeves and look at
Transactions and Isolation Levels.

CheckMessage

Do Until Company.Revenues < 0
MessegeID = broker.getNextMessage()
broker.sendMessage(MessageID)
Loop

SendMessage(MessageID)

Create a new thread...
SendMessage(MessageID) in this new thread...
Return to the MessageBroker.CheckMessage...while the other thread is
preparing and sending the message...

This will continuious poll the database server for new messages! ...

------------------------------------

However, if it is mission criticial users receive this information ASAP
... TRIGGER on database table! If the user can wait 10 seconds .. build
a BROKER ... and have it spawn as many threads needed to send the
messages in the QUEUE ... have it control the process...and sending!

------------------------------------

Again, many solutions, depends on your needs ... code sample, table
structure ... some thing to trying and figure out how exactly your
'threads' are getting 1 message...

Jeff.


"Jay" <someone@xxxxxxxxxxxxx> wrote in message
news:OLc3QDxEHHA.996@xxxxxxxxxxxxxxxxxxxxxxx
It is extremely important for the employees to receive the messages
almost immediately after identifying (time val in table). So, threading
ensures I can execute multiple processes at the same time... only
problem is these threads read from the same queue table and have the
potential to send the same message to the same guy x number of times.
where x is the number of threads running.

"jeff" <jhersey at allnorth dottt com> wrote in message
news:uQ7%23Y4wEHHA.4404@xxxxxxxxxxxxxxxxxxxxxxx

Still do not know why you need 4 threads ? Speed ? Does runnig the
process in one thread not work fast enough for you since timing is
import ? Why 4 threads why not 6,12 , 3, 36 ... ? How do you
determine which row a 'thread retrieves' ??? does it simply select the
next available row from the DATABASE ??? How does the thread determine
'the next row'???

If mutli-threading is duplicating 'send messages' it is not working
properly...
If you are multi-threading to improve performance ... maybe look at the
design...
If you need to implement a locking mechanism / or / logging mechanism /
or / a checking mechanism to avoid duplicate messages caused by
multi-threading ... these will all come at a cost ... performance cost!

What is bottle necking your process that you need 4 threads? Is it the
READ from the database ... is it the SENDING the text message? Is it
connecting to the database ??? do not know here ...

Look at what is causing problem ...

- you need better performance ... so, multiple thread it! However,
mutli-threading the entire process is causing issues ... duplicate
records ... now you either need to incorporate a locking procedure ...
or a checking procedure to avoid duplicate messages... all this has a
cost to the overall performance ... adding more threads may in fact
negatively impact the overall performance....

Maybe implement a message broker ...

- message broker gets all the necessary messages or message id's from
the database to be sent ... in-memory list...
- message broker loops through list of messageIDs ...
- message broker starts another thread for processing the
sendtextmessage functionality for each messageID...
- message broker will include the messageID so the process knows which
message to get and process...
- message broker will ensure each database row is only processed
once...

It is very hard to help you without knowing exactly what you are
doing...

Again, I will ask, what rational / reasoning did you use for using 4
threads ? Performance ? Speed ? ... Where is the bottle neck in the
process that requires you to multi-thread it? Maybe just move the
'bottle neck section' to another thread ... ?

Grasping at straws ...

Code would be nice how to see what you are doing ...

Jeff















"Jay" <someone@xxxxxxxxxxxxx> wrote in message
news:uBX7QewEHHA.1224@xxxxxxxxxxxxxxxxxxxxxxx
Thanks. I do need to select an individual row at a time and all 4
threads need to do this. What trasaction isolation level would you
recommend? Perhaps a stored proc may be faster to execute and return
the values as opposed to building the transaction in the code. What
has to happen is every 5 seconds, and for each thread, a sub runs to
get a single row then send the message to dtr("numbertosendto").
Because this app heavily relies on timing it is important that all
threads run and only one distinct row can be returned at a time for
each thread.



"jeff" <jhersey at allnorth dottt com> wrote in message
news:%23qVFIYwEHHA.3576@xxxxxxxxxxxxxxxxxxxxxxx

are you selecting individual rows at a time from the database table
...

if so ... use a transaction ...

begin transaction ...

select a row from database ...

update the row in table set flag = 'Processed'

end transaction

this will lock the row until the end of transaction is issued ... as
long as the isolation level is set accordingly...

Again, please let us know how you are getting the information from
the database... then we can help!

If you are reading a bunch of rows in one statement, storing the rows
in in-memory datasets on the workstation, looping through the rows
one by one ... then you may need to either ...

implement as above locking only the records you retrieve / update -
need to watch out here for table locking ... may impact performance,
implement using an update flag at sent and 'check before send
method'...
...
lots of options here, just need to know how you are retrieving your
data...and how you are processing the data.

Question, why do you need 4 threads running ... are they doing
'different processes', sending different 'types' of messages ...
sending the same message ... just need 4 to make it faster ... what
is the logic ...

Tieing a record to a thread my cause problems in the future ... what
happens when a thread stops or hangs ... those messages will not be
processed... When happens if somebody changes the ThreadID ... in the
program ...

trying to help ..

Jeff

PS: you can lock individual rows ... look at how database
transactions work and incorporate it in you program ...








"Jay" <someone@xxxxxxxxxxxxx> wrote in message
news:%23f70LJwEHHA.2328@xxxxxxxxxxxxxxxxxxxxxxx
Thanks for the reply.

I do need all 4 threads running (maybe even more in the future). I
can not delete the row from the table... it needs to be there for
later update use. I am considering marking each row at insert (1 -
4) then having each of the 4 threads only select a row based on the
mark. It's becoming a little tricky... it would be great if I could
lock a single row when selecting it.


"jeff" <jhersey at allnorth dottt com> wrote in message
news:OPCyvDwEHHA.4740@xxxxxxxxxxxxxxxxxxxxxxx

Question:

What is the reason for the 4 threads?

Do you need all these threads processing the same dataset?

Just asking, maybe you could change your approach to avoid this
problem...

If not, investigate Record Locking / Transactions...

How are you retrieving your records in to memory for processing ...

ie,
- are you selecting on the entire contents of a 'message' table and
looping through an in-memory dataset...

- do you retrieve 'chunks' or rows from the message table based on
parameters...

You could use the internal locking mechanisms of the database to
achieve this ... but procede with caution...

Start Transaction...
SELECT * FROM MESSAGE TABLE
...
stuff this into a data table on the desktop
...
DELETE * FROM MESSAGE TABLE
...
punt / clean the table.
...
End Transaction.

The transactions places locks on the table ... but selecting the
entire tables contents ... you are essentially 'LOCKING' the entire
table until the transaction is over...

So, each thread's 'retrieve' process will have to wait in line
until the select and delete are processed. As well, anything
triggering 'New Messages to be Added' will be delayed until the
Transaction is completed...

You may impact the overall performance of you application ... need
to investigate.

If locking does not work for you, you will need to implement some
type of logging / checking approach...

ie... have a log table...once a process has sent the message, write
to a log table ... message sent. Before each message is compiled
and sent, check the log table to see if another process has sent
the message...if not, send you message...


A snip of code would be great to figure out how you are retrieving
a list of messages and how you are processing each record...

Jeff.


"Jay" <someone@xxxxxxxxxxxxx> wrote in message
news:O2phzwvEHHA.4280@xxxxxxxxxxxxxxxxxxxxxxx
I have a multi threaded VB.NET application (4 threads) that I use
to send text messages to many, many employees via system.timer at a
5 second interval. Basically, I look in a SQL table (queue) to
determine who needs to receive the text message then send the
message to the address. Only problem is, the employee may receive
up to 4 of the same messages because each thread gets the recors
then sends the message. I need somehow to prevent this... just
can't think of how. Somehow I need the other threads to know that
another thread is already using that record and move on to the next
record. I thought of getting the record then marking it (column
value) as in use and writing the code to look for records only
where not in use... problem is all 4 run fast enough to all
use/mark the row. Any thoughts?

Thanks a lot.

Jay















.

Relevant Pages

  • Re: transactional replication problem
    ... All transaction from killed process were rolled back... ... Did they get deleted from distribution database ... >Looking for a SQL Server replication book? ...
    (microsoft.public.sqlserver.replication)
  • Maintaining Transactional Integrity with OPLOCKS ?
    ... We use an ISAM style database at the moment and suffer from clinet machines ... indexes stored on the server. ... This is a performance enhancement to the Windows client ... the transaction integrity of the database structures on the server ...
    (microsoft.public.win2000.networking)
  • Re: Win2k3 SP1 error: New transaction cannot enlist in the specifi
    ... What we came upon was the fact that Windows Server 2003 requires roles to ... Do you use SQL Username and password, or SSPI to logon to the database (if ... >> This dll supports transactions, but does not initiate transaction calls. ...
    (microsoft.public.windows.server.general)
  • Re: Synchronized Applications
    ... I believe SQL Server calls the first method you described "log shipping" (in ... combination with a transaction queue). ... read somewhere that the database is inoperative while it is importing logs. ...
    (comp.lang.java.programmer)
  • XmlHttp (AJAX) and postbacl
    ... is retrieving a table from my database and ... through post back the list i have created through javascript just ... Will i be able to send that list to the server? ...
    (microsoft.public.dotnet.framework.aspnet)