How to update database items using multiple threads within a transaction?

How to update database items using multiple threads within a transaction?

I hope you can help me
I'm writing an ASP.NET application and I have to loop on over more than
70,000 datarows within a datatable in order to update database records
(each datatable row has information to update a database record). My idea
was to break the table into 10,000 rows tables and create some async
methods to update the database (one async method per table). I need to
make sure every record on the table is updates, so everything is wrapped
into a TransacionScope block.
The problem is, when I run the Page.ExecuteRegisteredAsyncTasks() I get a
Timeout exception telling me that the max allowed time to get a connection
was finished.
My code is as follows
using (TransactionScope transaccion = new
TransactionScope(TransactionScopeOption.RequiresNew, new TimeSpan(0, 120,
0)))
{
//Validando momentos
clConsultoraCompensationPlan.asignaValoresFinales(dtConsultoraCompensationPlan,
compensation_plan, usuario,pagina);
transaccion.Complete();
}
private static void asignaValoresFinales(DataTable dtConsultorasMaster,
clCompensationPlan compensation_plan, int usuario,Page pagina)
{
List<DataTable> lista_tablas = new List<DataTable>();
//Separadno la tabla en 30 tablas
foreach(IEnumerable<DataRow> renglones in
LinqExtensions.Split(dtConsultorasMaster.AsEnumerable(), 50))
{
lista_tablas.Add(renglones.CopyToDataTable());
}
foreach (DataTable dtConsultoraCompensationPlan in lista_tablas)
{
AsignaValoresFinalesAsincrono tarea = new AsignaValoresFinalesAsincrono();
PageAsyncTask tarea_asincrona = new PageAsyncTask(tarea.OnBegin,
tarea.OnEnd, tarea.OnTimeout, new
ContenedorAsignaValoresFinalesAsincrono(dtConsultoraCompensationPlan,compensation_plan,usuario),
true);
pagina.RegisterAsyncTask(tarea_asincrona);
}
pagina.ExecuteRegisteredAsyncTasks();
}
In order to run the query I'm using a singleton instance, does this affect
something? Is there a better way to complete this task rather than split
the table and multithread?
Thanks in advance