The following are well known hybrid computing models:
z/OS Hybrid Batch is another model that integrates non-z/OS platforms such as Unix or Windows with z/OS batch. Many platforms are supported:
The z/OS Hybrid Batch model incorporates the traditional z/OS batch strengths of scheduling, automation, and access to enterprise data with non-z/OS platforms. The following diagram shows how Co:Z Toolkit components support the z/OS Hybrid Batch model. The Co:Z Launcher component resides on z/OS and the Co:Z Target System Toolkit is installed on each non-z/OS platform. In this example, z/OS jobs can access cloud servers as well as IFL engines.
OpenSSH is used to establish secure, authenticated network connections. By default, data transfer is tunneled (encrypted) over the ssh connection. Optionally, data can be transferred over raw sockets which offers very high performance without encryption costs, ideal for a secure network, such as zEnterprise HiperSockets.
Additionally, security using the z/OS Hybrid Batch model is z/OS centric. All z/OS resource access is through the job step, controlled by SAF (RACF/ACF2/TSS) and normal user privileges. Remote user credentials can be stored in SAF digital certificates extending the reach of the z/OS security envelope to the target system. Shared certificate access enables multiple authorized z/OS users to use a single target system id. Finally, Dataset Pipes streaming technology can be used to reduce "data at rest".
For a simple example (including JCL) illustrating how z/OS Hybrid Batch works, see the following ten minute video: Introduction to z/OS Hybrid Batch Processing . Co:Z Launcher starts a program on a target server and automatically redirects the standard streams back to jobstep DDs. The target program running on the remote server uses Co:Z Dataset Pipes commands to reach back into the active jobstep and access z/OS resources:
When designing a z/OS Hybrid Batch solution, consider the following principals for z/OS Hybrid Batch processing:
For examples illustrating the z/OS Hybrid Batch model in detail, see the following case studies: