Get Off the Data Center Floor: Dispelling Server Migration Myths
Author: Charles Wright, CEO, CTO, Workload Portability Evangelist at ATADATA
During the course of my career, I have spent innumerable days and nights on data center floors performing migrations and managing transformational projects. As a result, I have acquired a unique perspective on workload migrations while deploying almost all commercially available automation products.
Simply put, I’ve experienced the same frustrations as many of my fellow transformation architects and migration engineers. In fact, the limitations of these technologies inspired me to create our software and form the team that is ATADATA today.
But we certainly weren’t the first company to address these limitations. Let’s look back for a minute, at a point when the migration space began innovating. The early pioneers in this space created breakthrough technologies. Yet, successive generations of entrepreneurs and engineers, eager to get a slice of the pie, began misleading customers by overpromising on what were only marginal improvements to the promises of auto-migration.
Despite innovations in the migration space, and things have changed for the better, many things remain the same. The outcome is server migration has never been easier than it is today, yet in most cases, execution remains as complex as it ever was.
The risk, confusion and frustration of moving a few dozen, or one hundred live workloads can be significant. Multiply that by hundreds or even thousands and the enterprise client is faced with massive challenges in migration.
Add to the mix, multiple platforms – physical, virtual and cloud – OS upgrades, and lack of complete visibility into one’s IT infrastructure like server interdependencies, it would seem that there is no single discovery or migration solution, especially for enterprise. That would be the stuff of science fiction.
Over time, the unfortunate outcome of under delivered promises on early innovations produced a bad taste in the mouths of customers, leaving them skeptical and cautious of the server migration space… and, who’s to blame them? The challenge now, is to convince the marketplace that TRUE innovation actually exists. It isn’t elusive or just more, empty promises portrayed by slick marketing campaigns and cryptic architecture diagrams.
Moving live workloads, regardless of source environment, direct to any target and auto-provisioning is very real. This is what the ATADATA team and I have created with the ATAvision and ATAmotion modular migration solution.
I’m going to debunk the top 4 commonly held beliefs around migration:
Current Belief #1: Agent based software installations are necessary and the industry standard (even though most customers want agentless.)
Reality: Agentless solutions in discovery and migration exist and no software installations are necessary. Middleware can accomplish the task of monitoring and executing these functions without software installation.
Current Belief #2: Two stage migration is the “best practice” for moving live workloads.
Reality: Absolutely not. In fact, not only is a straight line the fastest between two points but it’s the most secure. Plus, it’s impossible to synchronize a highly transactional database over two steps via an image library. Single stage migrations have been the standard; it wasn’t until cloud API’s confused developers into thinking an extra step was required. Not true.
Current Belief #3: Shipping your database to the cloud on a truck (literally) is the only practical method of moving live workloads.
Reality: Fedex might be your “go to” for shipping boxes but it shouldn’t be when it comes to moving live workloads. Besides, it’s impossible to keep your databases synchronized if they’re packed on a plane or truck. Moving live workloads over the wire is not only secure as properly architected but allows you to maintain synchronization of your databases with no downtown or disruption to your business.
Current Belief #4: Asynchronous host level block syncing for OS and databases in the cloud, is current best practice.
Reality: A methodology that looks great in an architecture diagram, but spells disaster in execution. There are well distributed case studies that cover the issues of block level replication. Keep looking if someone tries to sell you on an innovation from 10 years ago reborn as something elegant for cloud applications... bottomline, it isn’t as easy as it’s made out to be. Application and OS consistent replication is the best practice.
In a nutshell, commonly held beliefs in automated migration no longer apply – even for the enterprise – innovation is real and TRUE live workload portability is a reality.
I have spent the last two years reaching out to my fellow migration architects and engineers telling them help is coming and now I’m telling them, and you, it’s here.
When you have a company whose engineers and architects live and breathe enterprise transformation, the result is an enterprise level migration software for workload portability that isn’t just a tool – it’s a solution.