July 22, 2024


This structure consists of the next parts:

Shifting knowledge from IBM z/OS to Google Cloud is simple with the Mainframe Connector, by following easy steps and defining configurations. The connector runs in z/OS batch job steps and features a shell interpreter and JVM-based implementations of gsutil, bq and gcloud command-line utilities. This makes it attainable to create and run a whole ELT pipeline from JCL, each for the preliminary batch knowledge migration and ongoing delta updates.

A typical movement of the connector consists of:

  1. Studying the mainframe dataset

  2. Transcoding the dataset to ORC

  3. Importing ORC file to Cloud Storage

  4. Register ORC file as an exterior desk or load as a local desk

  5. Submit a Question job containing a MERGE DML assertion to upsert incremental knowledge right into a goal desk or a SELECT assertion to append to or change an present desk

Listed below are the steps to put in the BQ MainFrame Connector:

  1. copy mainframe connector jar to unix filesystem on z/OS

  2. copy BQSH JCL process to a PDS on z/OS

  3. edit BQSH JCL to set web site particular setting variables

Please check with the BQ Mainframe connector weblog for instance configuration and instructions.

BigQuery is a totally serverless and cost-effective enterprise knowledge warehouse. Its serverless structure enables you to use SQL language to question and enrich Enterprise scale knowledge. And its scalable, distributed evaluation engine enables you to question terabytes in seconds and petabytes in minutes. An built-in BQML and BI Engine allows you to analyze the information and achieve enterprise insights. 

Dataflow is used right here to ingest knowledge from BQ to Elastic Cloud. It’s a serverless, quick, and cost-effective stream and batch knowledge processing service. Dataflow supplies an Elasticsearch Flex Template which may be simply configured to create the streaming pipeline. This weblog from Elastic exhibits an instance on configure the template.

It is attainable to load each BigQuery and Elastic Cloud totally from a mainframe job, without having for an exterior job scheduler.

To launch the Dataflow flex template immediately, you may invoke the gcloud dataflow flex-template run command in a z/OS batch job step.

In the event you require extra actions past merely launching the template, you may as an alternative invoke the gcloud pubsub matters publish command in a batch job step after your BigQuery ELT steps are accomplished, utilizing the --attribute possibility to incorporate your BigQuery desk identify and every other template parameters. The pubsub message can be utilized to set off any extra actions inside your cloud setting.

To take motion in response to the pubsub message despatched out of your mainframe job, create a Cloud Construct Pipeline with a pubsub set off and embrace a Cloud Construct Pipeline step that makes use of the gcloud builder to invoke gcloud dataflow flex-template run and launch the template utilizing the parameters copied from the pubsub message. If you must use a customized dataflow template slightly than the general public template, you need to use the git builder to checkout your code adopted by the maven builder to compile and launch a customized dataflow pipeline. Further pipeline steps may be added for every other actions you require.

The pubsub messages despatched out of your batch job may also be used to set off a Cloud Run service or a GKE service through Eventarc and may be consumed immediately by a Dataflow pipeline or every other software.

Mainframe Capability Planning

CPU consumption is a significant component in mainframe workload price. Within the primary structure design above, the Mainframe Connector runs on the JVM and runs on zIIP processor. Relative to easily importing knowledge to cloud storage, ORC encoding consumes far more CPU time. When processing giant quantities of knowledge it is attainable to exhaust zIIP capability and spill workloads onto GP processors. You might apply the next superior structure to cut back CPU consumption and keep away from elevated z/OS processing prices.


Source link