By: Joe Panettieri
I have a mantra in my business: if you can't automate it, then don't do it. In other words, our IT team doesn't begin a journey if the path to success (and ongoing maintenance) depends on manual processes. The reason? As our business grows and evolves, a growing mound of manual processes could consume time that's better spent on innovationโparticularly DevOps.
But what does true automation look like today? And how will it evolve over the next five years or so?
For the most part, the bulk of today's infrastructure automation involves (1) workload management. But if you look down the roadโstretching from today through about 2020โwe'll see more and more demand for (2) cloud and virtual system automation; (3) self-service tools; (4) orchestration; and (5) brokering solutions, notes IDC. I agree.
Five areas to watch
Let's take a closer look at each one of these five areas, their potential benefits, and the related challenges.
1. Workload management: Here, traditional job scheduling software is evolving into workload management and automation software. It typically involves:
- Batch processing, which takes a workload and executes it at a specific date and timeโsort of like only driving on a highway when traffic is lighter than usual.
- Event-driven process automation, which means the process happens in reaction to a specific eventโperhaps a specific set of patches from a vendor arrives and is available for activation.
- Service-oriented job scheduling, which means applications communicate with one another over a networkโcalling on each other for updates without human intervention.
Dating back to big iron and Unix systems, the workload management market is quite matureโthough new technologies for x86-centric data centers continue to arrive and push the market forward, growing it at low single-digit rate from about 2015 to 2019, IDC expects.
2. Cloud and virtual system automation: Here, IT administrators can move workloads between private, public, and hybrid clouds. Ideally, the associated tools offer unified security, governance, and compliance across applications, as well as physical and virtual infrastructures.
Cloud and virtual system automation are critically important to the DevOps movement. Each time developers release new code or new applications to IT operations, automation can help speed the software's rollout across infrastructure. It's sort of like giving your software an approved "passport" to move from one network to another before settling in and activating itself on the most appropriate virtual or physical server.
3. Self-service: Many end users know how to manage SaaS workloads in public clouds. Now, that capability is increasingly coming to private and hybrid clouds and digging down to the infrastructure level.
Imagine, for instance, if users could manage their own high-performance computing (HPC) workloads. Many businesses are making that a reality today.
Indeed, self-service portals allow DevOps, end users, and big data experts to "request infrastructure and application resources, submit jobs, and view the progress, estimated completion time, and results through a user-centric visual dashboard," HP's Helion Self-Service experts have noted.
4. Orchestration: This is a biggie, and it's also going mainstream now. Orchestration is the coordination and management of tasks across IT incidents, changes to services, system-generated events, and service requests, as well as routine and ad hoc activities.
Orchestration's three big benefits: First, the potential for human error is reduced by predictable, repeatable, and automated processes. Second, it potentially cuts your maintenance costs by eliminating manual, repetitive tasks. And third, it cuts the time you need to deploy new infrastructure.
5. Brokering solutions: Here, the IT department transforms into a service provider for end users. It's still early in the brokering game, but the goal is for IT departments to offer automated governance, chargeback/showback, service catalogs, and app stores to employees.
DevOps, for instance, could receive chargeback information for consuming IT operations' time and compute resources for specific projects. Similarly, marketing could potentially access a catalog of services to promote a new product to customers. As the marketing department scales up or down specific promotional workloads running on IT infrastructure, the marketing department would receive and pay the resulting chargebacks.
In essence, most pundits believe the IT department will become a cloud services provider (CSP) and broker to employees. The big, big question remains: When?
The view from 2020
By the start of the next decadeโperhaps even soonerโIT infrastructure should transform into an on-demand utility, similar to mainstream utilities such as water, electricity, phone, and Internet services.
You never see all the automated maintenance, but that infrastructure pumps water, electricity, phone, and broadband service to end users at an incredibly reliable rate. Some end users pay for the utility as they go (that is, the chargeback system). Others pay a flat monthly fee, such as for broadband.
We've already seen similar models take hold in the IT market. But the push toward enterprise brokering solutionsโthat is, total automation for end usersโis the next wave to watch.
Learn more about the DevOps journey and decide if this mindset could help your organization.
Content czar at After Nines Inc., providing strategic IT guidance from 9:01 daily. As a journalist, analyst and IT media entrepreneur, Joe has tracked enterprise IT issues since 1992.
Connect with Joe:
