IT process automation sucks, is expensive and doesn't work . . . NOT!
- 26 February, 2013 18:02
This vendor-written tech primer has been edited by Network World to eliminate product promotion, but readers should note it will likely favor the submitter's approach.
It's commonly perceived that IT Process Automation (ITPA) costs too much and - after a painful implementation process - doesn't work. But enterprises taking a "crawl, walk, run" approach with Automation as a Service (AaaS) are realizing dividends.
These users avoid expensive and complex enterprise software platforms, upgrade tactical scripting approaches, and reject "Free" ITPA bundle offers from large software platform companies. Whether you are ready to adopt the unfortunately named, but high ROI, Automation as a Service (AaaS) approach, or pursue other ITPA models, here are a few considerations to keep in mind.
One thing to keep in mind is you get what you pay for. Although ancient advice, it still makes sense, especially in today's confusing world of enterprise software applications and IT process automation. As you approach your "crawl, walk, run" to ITPA, consider selecting a platform that scales. Plan for success so when you start "running," your gear can carry you the distance.
Large vendors like Microsoft bundle ITPA solutions for free. If the bundled solution were truly valuable, they would be selling their ITPA solution and bundling something else. Conversely, by selecting the right automation software, the software cost pales in comparison to business benefits from the automations implemented.
In lieu of free software, consider a "pay for use" model rather than paying upfront fees for enterprise licensed software. Explore whether security and compliance requirements allow for a SaaS (software as a service) implementation so you can roll out your ITPA as a SaaS solution and bring it in-house later, if appropriate - and if your chosen vendor allows such flexibility.
Whatever platform, test and pilot first. You will learn what is most important for future scaling needs. If the POC works well and embeds the capabilities you will require for RUN-mode operations, then you're set for the future. If not, you are now informed as to what is needed and can make a better selection for your next implementation.
Success may not require an Automation as a Service (AaaS) approach, but the right choice, early on, helps enable efficient and effective scalability.
If an AaaS solution seems like a good fit, consider process complexity based on the number of steps involved and the number of system interfaces. Vendors may charge based on steps, interfaces, or a combination. Start with low-risk processes measured by end user perceived importance and use frequency. Get comfortable with how the platform works, then scale up.
During evaluation, check if the platform can be brought in-house in the future. This option may be required for compliance, cost or performance. Some vendors are adamant about only providing a SaaS solution; others are more accommodating. Sometimes, you'll never need to bring it in-house, but we've seen situations where global implementations with stringent performance requirements require behind-the-firewall deployment after the initial SaaS implementation proved successful.
Finally, in evaluating an AaaS Service offering, confirm whether the vendor provides a managed service wherein they operate and maintain the automation to accommodate environment changes - systems upgrades, work flow improvements, for example - sometimes called sustenance services. A base offering of monthly hours in a managed service offering is allocated for client requests, "tweaks" or minor modifications to ensure automations remain relevant.
Whether you are attempting ITPA for the first time, are a more mature organization scrapping existing approaches for a better approach, or are considering trying a new or additional path forward, let's explore "crawl... walk... run".
* Crawl. Start small, with one process - a low-risk, non-mission-critical process. Don't tackle complete server builds and patches when you could start with server decommissioning. It may seem obvious to start slow, but many organizations jump headlong into automation only to get into trouble, requiring serious and costly assistance.
Be clear on the desired process flow, outcomes and service-level agreement performance metrics. What benefits will be seen upon automation completion? Developing business cases (even a mini-version for smaller jobs) is another frequently overlooked task. Make sure it happens!
It's a pain, but document processes step-by-step. Run documentation by colleagues and end-users to capture all questions, exceptions and feedback. The "how to" may seem to be already integrated into Standard Operating Procedures (SOP), but there's no substitute for a fresh process review and validation.
* Walk. There may be pressure to get more done faster, but it often results in a mess. It's critical to communicate your plan so stakeholders see the finish line. Hold them at bay when pressure builds to just "get it done" as you head toward that finish line.
Maintain your pace and focus on winning the whole race, not just one leg. Breakdown implementation steps in an orderly fashion to get from the current to the desired state. A certified project manager, process optimization expert, or process automation designer can be valuable in leading this effort. Often, vendors will provide this service or know experts that fit your budget. Carve out the required staff and resources throughout the forecasted timeframe to complete the project.
With an implementation plan in place and resources assigned, it's time to build process steps, interfaces and workflow into the automation platform and begin test runs to evaluate that flow. Ideally, have a test environment, if not, test in your operating environment during maintenance windows until you are certain the newly designed process meets objectives and is stable enough for production-grade operations.
Ensure a means for sustaining the automations over time when future systems or process changes occur. Too often, the focus is entirely on deployment and the team lacks the foresight or continued energy to ensure ongoing deployment sustenance.
Integrate change management processes to accommodate new automations before production implementation. Don't accomplish this task and be prepared for the classic "two steps forward / one step back" problem. You will be stuck with automations breaking down upon inevitable environment changes.
* Run. OK, you've exercised your muscles and have a success under your belt. It's time to stretch. Extend automation initiatives to new applications - one-by-one. Over time, implementing multiple processes or linking them together is do-able. Beware of moving too hastily, before securing some solid wins. Once, you've got a few under your belt, let 'er rip!
Tackling ITPA infrastructures according to these basic guidelines has proven successful for a number of companies. The tips sound simple, but it's surprising how many knowledgeable enterprise IT organizations miss them and suffer the consequences of rework, delays, process breakdowns, and more.
IT automation doesn't have to suck. It doesn't have to cost an arm and a leg or put anything in a sling! You might want to try Automation as a Service or some of these other tips to protect your "AaaSets."
# # #
About the author:
Prior to joining Appnomic Systems, Mr. Solnik was the president & COO of OpSource (a Silicon Valley Infrastructure as a Service (IaaS) company recently acquired by Dimension Data, a $6 billion global systems integrator) where he was responsible for the operations of over 200 web applications including some of the most successful SaaS and software apps in the world. His customers relied on him and the OpSource team to ensure their applications were performing well - 24x365. After spending a number of Saturday nights resolving customers' application operations issues with the OpSource team, he decided to join the vendor community and bring better solutions to prevent operations issues in the first place.
Read more about infrastructure management in Network World's Infrastructure Management section.
Join the Computerworld Australia group on Linkedin. The group is open to IT Directors, IT Managers, Infrastructure Managers, Network Managers, Security Managers, Communications Managers.
AusCERT 2013: NBN users need security professionals’ help, says Google
Despite poor work-life balance, IT pros like their jobs: survey
WikiLeaks Party closer to registering
WikiLeaks Party closer to registering
Telstra COO reveals restructure