By Taos
Some robots have a bad reputation (think Terminator), but let’s be honest. There are many things robots do better than we corporeal beings. “And what might those be?” you ask. According to The Robot Report¹, the top 5 things robots do better than humans are:
- Handling tedium
- Extreme sensing
- Strength and speed
- Unwavering focus
- Perfect, objective recall
When applied to computing infrastructure and operations, this list — with the possible exception of strength — is very compelling. Computers are robots, after all, so their use in managing other computers is not only recursive, it makes life easier for us non-robots. Let’s look at how.
Handling tedium — Anyone who has been responsible for systems monitoring and alerting knows about ‘handling tedium’. (The phrase “months of boredom punctuated by moments of terror²” is an apt description.) Simply put, the ongoing management of large-scale IT environments is mind-numbing for mere mortals.
Extreme sensing — Within the context of systems operations, sensing involves the proper detection, classification, and response to events. Failure to do this properly results in type I and/or type II errors³. False positives mean you’re waking someone up in the middle of the night for no reason. Frequent false negatives likely mean you’re updating your resume or considering a career change.
Speed — Whether transferring files, crunching numbers, or querying databases, computers can process far more information than even the smartest human Jeopardy player.
Unwavering focus — Unlike the roving hordes of drivers, pedestrians, and babies who are increasingly dependent on their smartphones for the essence of life, computers are good at both multi-tasking and completing assigned tasks.
Perfect, objective recall — Computers are built to store and retrieve information without regard to political, religious, or peer pressure. By contrast, human memory is fallible in far too many ways.
My point in praising robotic capabilities isn’t to make people feel inadequate by comparison. Rather, it’s to point out that computer systems management — whether on-premises or cloud-based — benefits from automation using other computers.
Freedom through automation
Because information technology has become a sine qua non of modern life and commerce, it’s increasingly important to deploy and manage it efficiently and reliably. If, like me, you’ve been involved in computer operations for more than a few years, you may have spent your share of time performing manual tasks in computer labs and data centers. From racking-and-stacking servers and storage arrays; to wiring up network and systems interconnects; to installing operating systems, applications, and patches; to performing system break/fix and hardware refresh; there are a lot of tasks involved in computer operations. When done manually, these tasks take a substantial amount of time and are error-prone.
Is it any wonder, then, that cloud computing is rapidly displacing on-premises data centers? With its promise of on-demand self-service, broad network access, and rapid elasticity⁴, cloud computing offers a better mousetrap. Similarly, one of the primary goals of DevOps — the marriage of software development and systems operations — is to enable “infrastructure as code⁵”? The robust growth of DevOps⁶ confirms the value of this capability.
Automation of low-level systems and functions frees developers and administrators to focus on higher-level capabilities including the non-functional requirements, or “ilities⁷” — availability, extensibility, supportability, etc. — required for true mission-critical systems. Rather than working directly with hardware, users work through a management interface that hides much of the underlying plumbing and presents consumable technology stacks and services.
This is what cloud service providers do. Take Amazon Web Services (AWS), for example. The last word in their name is ‘Services’ — not hardware, not software, not technology. Using the AWS console user interface (UI), application programming interface (API), or command-line interface (CLI), users can provision the services they need to create higher-level services and applications.
Living in a hybrid world
While vendor-specific tools are very useful when working within their respective technology stacks, most organizations live in a hybrid world. They may have some workloads deployed on-premises and others in one or more clouds. Rather than using separate UIs and APIs for each, a cloud management platform provides a consistent interface across all their operations.
This is what the Taos Management Platform (TMP) does. It serves as a unified interface for a variety of functions including:
- Configuration management database (CMDB) for managed systems and components, including on-premises and cloud
- Ingestion and alerting platform for systems monitoring
- Service catalog of templatized, approved configurations
- Unified security and governance monitoring, including execution venue control
- User request and self-service portal
- Quota management and spending approvals
- Ability to create complex workflows and orchestrations across components and environments
- Cost tracking, trending, alerting, and optimization recommendations
- DevOps enablement through cross-environment provisioning API
- Near real-time usage, trending, and exception reporting
In short, TMP serves as the manager-of-managers ‘robot’, providing control and oversight across all managed workloads. Best of all, access to all TMP functionality is a standard part of Taos’ managed services.
For Further Reading:
[1] https://www.therobotreport.com/5-things-robots-better-humans/
[2] The New York Times Current History of the European War (1915), Google Books
[3] https://en.wikipedia.org/wiki/Type_I_and_type_II_errors
[4] https://csrc.nist.gov/publications/detail/sp/800-145/final
[5] https://en.wikipedia.org/wiki/Infrastructure_as_code
[6] https://www.grandviewresearch.com/press-release/global-development-to-operations-devops-market
[7] https://en.m.wikipedia.org/wiki/Non-functional_requirement