I have a requirement where I have to access the cloud database thru mainframe. So to do that we planned to use MQueue series.
It took around 5 min to process 100 records to send the request to MQ and receive the response in the response queue. We process 1 record at a time meaning ( open all the connection /queue in the beginning and then we send the 1 rec of primary key to the MQueue then we receive the rec corresponding to that primary key in the response queue and do the same process for 2nd record till 100th record in the end close all the MQ connection) is there an optimistic way to reduce the time.
Or is there any way to unload the entire cloud database and populate in the MQ or you have any other suggestions
We got a new requirement to generate message for new client and currently all the data’s are available in cloud. From mainframe we have no idea how to connect thru cloud so we decided to use MQ to do request and response from cloud database. No people from here have any idea how to interact cloud thru mainframe?
My client doesn’t want to clone the cloud database to DB2 as the cloud database is used by many applications currently. So we have to use the cloud database only for sure. Now we are trying to do performance process either to unload the cloud table and copy to a temp table or to process directly thru MQ.
Joined: 20 Oct 2006 Posts: 6970 Location: porcelain throne
other than the obvious
- your company pay for some training of the people that is the basis of the billing of your customer..............
you mentioned mqs, but i don't think that you are connecting directly to cloud with that.
cloud is accessable thru internet
(more accurately, cloud is based on servers)
which means you are using mqs to communicate between the mainframe
and some server based something that deals directly with cloud.
i would stop trying to find fault with mqs
mqs is really quick
and unless the task priorities on the mainframe and your server are being set by neanderthals,
your hold-up is not mqs.
maybe the data that you are transfering is just too big.
try smaller packets until you learn and understand what needs to be done.
you provided very little info.
what kind of response do the applications that currently access the cloud db have?
also, you have to generate a message on the mainframe, triggered by something,
why not have that something also include the info from cloud that is necessary.
i realize that would require a complete revamp of the process,
but all you company did was spit in the wind
and now it is smacking you in the face.
designing something without understanding your db is asking for trouble.