Red Hat presented a vision during today's keynote address at the Red Hat Summit that it believes Ansible can and will be the foundation for enterprise-wide automation. Red Hat, by the way, recently acquired Ansible.
Although the vision appears enticing at first glance, the broad statements about unifying the enterprise data center under Ansible really don't ring true.
What is Ansible?
Ansible describes its technology in the following way:
Ansible is the most popular open source automation tool on GitHub today with more than a quarter million downloads per month. With over 2,400 contributors submitting new modules all the time, rest assured that what you are automating is covered in Ansible already, or will be very soon."
Ansible was founded to provide a new way to think about managing systems and applications that better fit this new world. Historically, management vendors and home-grown scripting solutions were created to manage stacks of software on servers. In contrast, Ansible was created to orchestrate multi-tier applications across clouds. From configuration to deployment to zero-downtime rolling upgrades, Ansible is a single framework that can fully automate today’s modern enteprise apps.
OUR DIFFERENCE
Simple, agentless & powerful. Ansible’s simple, YAML-based automation syntax is quick to learn and easy for the entire organization to adopt. By utilizing SSH (on Linux/Unix) or WinRM & PowerShell (on Windows), Ansible doesn’t require an agent to do its magic. And Ansible does so much more than configuration management… Ansible can deploy apps and orchestrate complex tasks that other systems have a hard time doing.
Snapshot analysis: Third parties must be involved
While an interesting thought, Ansible, or for that matter any X86-based management and automation platform isn't sufficient to take on the role of enterprise-wide automation and management platform. In a traditional enterprise data center, there are just too many non-X86 platforms and solutions.
Providers of those other platforms, such as mainframes, midrange Unix systems, midrange single vendor system, and other non-X86-based platforms, must be persuaded to do part of the work to integrate their platform with Ansible. Another possibility, but one that wouldn't be supported by the primary supplier, is that the Ansible open source community does the integration work to allow Ansible to provide management and automation of workloads.
What is far more likely is that Ansible, at best, be only a part of a comprehensive solution for enterprises.
The enterprise data center looks like a computer museum. Today's workloads are built upon the foundation of those built 10 years ago. Those workloads are, in turn, built upon the foundation of those acquired or developed before that. If one wanders through the racks of many enterprise data centers, one is likely to find:
- Mainframe applications acquired or developed in the 1960s through today
- Unix applications executing on systems from IBM, Sun (Oracle), HPE (DEC, Tandem, HP), Hitachi, NEC and perhaps many more
- Single-vendor environments, such as IBM I (successor or System 3, System 34, System 36, etc.)
- Industry-standard X86 systems running Windows, Linux, Netware and Lord knows what else
- Intelligent networking equipment from a host (pun intended) of suppliers
- Intelligent storage equipment from a host (pun also intended) of suppliers
- Intelligent power management equipment
- Intelligent cooling equipment
- An amazing array of other things, such as facilities security, telephone systems, etc.
This means that when a supplier, Red Hat in this case, presents that an X86-based system is prepared to take on the role of the "one ring," the ring to rule them all, it is not likely to be completely true.
At best, it might help unify the management and automation of the industry standard workloads and, possibly, workloads on some midrange Unix systems. It is unlikely to help enterprises manage everything in the data center.
I'm looking forward to having a conversation with Red Hat about this after Summit ends.