Changing engines in mid-flight

Speaking from the control tower, it is difficult to muster enthusiasm for a risky migration when an organization is storming and norming. A corporation is the sum of its parts, especially true when it is the result of mergers and acquisitions. Imagine the people, process, and tools unity that needs to occur. This is bigger than Brexit - well, figuratively speaking. Transformation is challenging amidst Consolidation. Yet imagine being presented with the funded initiative for just such a migration - to consolidate from four IT transaction systems to two, within the fiscal year. As though that were not enough, bring monitoring into the 21st century, is the mandate from above. Move over ancient agent-based monitoring systems that take a decade to deploy, even longer to configure. Usher in AIOPs with event correlation for pin-drop silence in environments that grow increasingly loud with incessant alerts in the eye of a perfect storm. The multi-colored post-its on a the most massive refrigerator could not cover the risks - and, yes, the rewards of such a multi-faceted make-over. Gravity is not the best friend of such a corporation in flight. Yet migration is a must.

First, there is the adoption and messaging phase. Stakeholders are up in arms due to decades spent on the well-worn, sagging, but much loved, and dreadfully obsolete comfy couch. This is the favorite Frankenstein of an IT Service Management system that allows you to customize every unnecessary but accustomed process in the corporate play book. The client has change freezes in every nook and cranny of the calendar and couldn't care less about boarding a more modern and fuel efficient airplane so long as it is not a **** 787. As for reporting, Smart Reporting be damned, they want their operational reports including exotic and unavailable, inaccessible fields in the database, extracted and served up every Monday morning on a silver platter with fresh smelling aromatic dark roast coffee. If it doesn't affect the sticker price, and so long as it is safe, so long as the end-client of the customer is catered to, who cares about changes in the menu aboard the aircraft? What then?

It is important to advertise the engine, emphasizing the features of interest and correlating to client bottom-line, security, efficiency, and - digital transformation. How does a migration pave the way to the holy grail of user experience - the coveted catalog. It is critical to address point of impact - after all that is where the rubber hits the tarmac. The blessing and the curse of a migration ultimately sits with the ultimate end-customer, the shopper in a smart shelved grocery store (I read of the Microsoft Kroger alliance yesterday). It rests with the health care customer who is entering a prescription refill. It resides in the palm of the mobile federal government official who is submitting a private ask for a protected server sitting somewhere in the undecipherable maze of a pentagon server farm. Sing the song of seamless user experiences, with faith in that of which you sing. Before going away for the week-end and meekly accepting the adage that everything will be fine on a Monday, test, test, test, with an eye on business user acceptability for the most basic and the most complex of workflows. A new technology willfully misunderstands the most rudimentary of user behaviors - acceptance born of habit. User Acceptance Testing is often unconscious because the customer expects to care less, and expects the provider to pay attention, even intuitively divine what matters most to the client. After all, the client is either footing the fuel bill to fly that metaphorical Migration airbus to Bangkok, or expecting a smooth, safe, seamless flight for free since the cost is covered in the contract. It takes a technical diplomat to be a migration ambassador (that could be a coveted title). It takes an "influencer" to urge the user community to test end-to-end. It takes a bit of a prophetic soul, with a penchant for forecast to anticipate the downstream impact of a migrated workflow. This is mostly unavoidable during a technology migration, or even during the switch of delivery model, from dedicated to leveraged, for example.

Automation and containerization are great allies of a convenient cloud migration. Thanks to Kubernates, Ansible, Azure, we are better together in an age of automated flight with cruise control. How great to be able to leverage a security-minded enterprise-ready configuration, assuming data likeness, compatible formats, and symmetrical tables. Truthfully, none of these can be taken for granted. While, technologically, most problems are not insurmountable, business cases come with constraints and the clock keeps ticking. To land the plane safely, there is great comfort in having the vendor as co-pilot. Before even embarking on the journey, the basics of such a trusted relationship must be cemented. Even if one is promised high class professional services, inevitably, the teams, distracted by the fire drills of daily operations, find it almost impossible to swing onto a moving plane that is moving at the speed of light. At some point capacity curbs enthusiasm for innovation. That is when two heads are better than one to navigate a flight path that can prove precarious.

Monitoring is a special air craft of its own. It is easy to earn consensus and gain approvals from a client community that has grown exasperated with the inefficiencies of legacy monitoring tools. However, it is equally dangerous to charge down a previously trodden path expecting the next best monitoring tool to be Excalibur promising complete clarity of vision and the cure for all ills. Truthfully, there is no single pane of glass that can exhibit an entire globe lit up with trouble spots showing say, network bandwidth utilization in red dots glowing loud and luminous. There is no absolute truth without thresholds, and those are tricky to align on. There is no glory in configuring clusters of devices of all arrays and types without a negotiated baseline. Success criteria crafted with customer consensus are especially important during a migration that could quickly turn contentious.

Migration is such a live wire. The expectation is that the tool will transform. The reality is the business knows their pain points the best. One must first crawl to configure basic thresholds for mission critical servers, exclude as necessary, conscientiously calibrate based on function- before one can walk, then run, then fly. What if an engine turns out to be much slower than the turbo-charged marvel one expects it to be, as sold (oops told). Can the other engine compensate? During the initial moves at the crest of the first wave of migration, what if one experiences that dreaded slowness due to a query that eventually brings the system down? It is the vendor who can rush into the fray and keep the migration vehicle afloat while looking at the underbelly of the plane, if you will, in an imagined space walk. Bring vendor and customer with you if you must migrate in a time of merger - all three co-pilots in a crowded cockpit. And then, hope for mostly fair weather and safe landings.