Operations

Steven J Zeil:

Last modified: Nov 30, 2020

Contents:

Abstract

Operations is what happens after the developers (think that they) are done with the software.

Fundamental components:

Deployment: setting up the servers that will run the software and installing the software on them.
Testing: conducting system/acceptance tests, possibly in conjunction
with the developers.
Operation: running the software and keeping it running.
Monitoring: measuring the performance of the running software.
Tuning: adjusting the software and/or the execution environment to to improve performance.

Is the software a service? (The developing organization runs the software.)

Includes web services and applications developed for internal use.
Need to set up the servers/machines on which the software will run.
Need to install the software and any required support packages on those servers.
Need to configure firewalls, access & security features.

Is the software a product? (We sell/lease it to others to run.)

Still need to system test in the anticipated operating environment(s).
Need to set up servers providing those environments.
Need to install the software and any required support packages on those servers.

If the software product is a service, we need to run it on our servers.

Monitor to ensure availability.
Recover from crashes (of software, hardware, network, etc.)
- Recovery may require repairing damaged files, databases, etc.

Do we need more servers?
- assuming the software was designed to be run simultaneously on multiple servers
  - No conflicts due to shared access of common resources (e.g., account database).
Should we reroute some traffic from heavily loaded servers to ones with lighter loads? ( load balancing )
Does a slow-running server need more memory? More disk storage?
Can we split or redistribute the components among the servers? (e.g., separate the web interface from the database manager).