Real distributed data access model with ado.net ?

[Guest]

Hello there,

I hope this is the right newsgroup to post this question...

Let's say I have a 3-tier .net app:
tier 1: sql server 2000
tier 2: win app client
tier 3: a web service delivering data from t1 to t2 using a DataSet.

The data transferred is of about 65.000 rows. This is a lot of data to
deliver with each client request.
What I want is to make the web service a bit smart so as to:
- return all the rows on the first client request
- return only rows with changes on subsequent requests
- return nothing if there are no changes at all

Any solutions / suggestions / references ?
Can web services achieve this kind of logic and functionality ?

Thank you.
Dimitris

 

[Miha Markic [MVP C#]]

Hi Dimitris,

Client should pass some kind of information of the last retrieval (timestamp
perhaps).
This info should be used by webservice when retrieving data.
However, you'll have problems seeing deleted data...

 

[Guest]

Hello Miha and thank you for your answer

I understand the logic of what you suggest but can you please tell me
exactly how the timestamp will be used ? As a parameter, in ado.net calls
maybe ? or as an extra column in the dataset transferred ? I am not quite
sure...

I would very much appreciate it if you could give me an example.

Thank you
Dimitris

Hi Dimitris,

Client should pass some kind of information of the last retrieval (timestamp
perhaps).
This info should be used by webservice when retrieving data.
However, you'll have problems seeing deleted data...

--
Miha Markic [MVP C#] - RightHand .NET consulting & development
SLODUG - Slovene Developer Users Group
www.rthand.com

Hello there,

I hope this is the right newsgroup to post this question...

Let's say I have a 3-tier .net app:
tier 1: sql server 2000
tier 2: win app client
tier 3: a web service delivering data from t1 to t2 using a DataSet.

The data transferred is of about 65.000 rows. This is a lot of data to
deliver with each client request.
What I want is to make the web service a bit smart so as to:
- return all the rows on the first client request
- return only rows with changes on subsequent requests
- return nothing if there are no changes at all

Any solutions / suggestions / references ?
Can web services achieve this kind of logic and functionality ?

Thank you.
Dimitris

 

[Miha Markic [MVP C#]]

Hi Dmitris,

Hello Miha and thank you for your answer

I understand the logic of what you suggest but can you please tell me
exactly how the timestamp will be used ? As a parameter, in ado.net calls
maybe ? or as an extra column in the dataset transferred ? I am not quite
sure...

I would say both. You should read timestamps for each table/record into
dataset.
When client wants fresh data, it has to pass Max(timestamp) for each table
that it is interested in.
I didn't ever try this approach thus I don't have any sample (reading
timestamp is trivial - never tried passing it back to sql server).
For seeking max time stamp from dataset take note, that timestamp is an
array of bytes.