I was driven to write this blog while considering the difficult choices so many IT leaders must make. Difficult choices made even more complicated by the combination of massive technology trends and a global pandemic occurring simultaneously. The migration decision is fraught with considerations around data protection, staffing, organizational design and business priorities, among many other factors, not the least of which might be cost.
Don’t get Choked up about the word “Cost”
I’ve spoken ad nauseam about cloud adoption strategy and how to make value judgements not based on “doing it cheaper” but on business and IT operational and execution benefit. Anyone who’s been bored enough to follow my LinkedIn or Twitter discussions will have been exposed to my gentle vitriol regarding cloud adoption not being about cost. That being said, cost is and always will be a factor.
So what’s the point?
Many of us have been caught in rather compromising positions during this pandemic and are now trying to play catch up. It’s also true that we are (or should be) in an on going battle to prioritize our portfolio projects. How do you weigh keeping an application in a colocation or on premises data center vs. cloud when there is also Edge to consider? How do we effectively leverage our limited funding to focus on those projects that will most effectively move the needle for our organizations? These are a few of the topics I want to touch on here.
Migration with data or without data, that is the question
Migration to a public cloud provider is a choice of platform and with that comes a balancing of risk and reward. Many have balanced risk by having some applications in public cloud and some in private cloud, others have also made efforts to use multicloud to solve for perceived or real performance benefits associated with a particular application or application environment. However, in an age where the sand is shifting ever more swiftly beneath our feet, how do we protect ourselves from being painted into a corner and more specifically a data location corner?
Considering Data Location in the Era of Edge Computing & Business Transformation
Gartner says that 75% of data will be created outside the cloud or on premises data centers by 2025. Business Transformation by virtue implies better use of data to make the business more agile, efficient and customer responsive, among many other things. The pandemic has illustrated for many that having a single location or fragile staffing and operations model can be a real risk to business continuity. All of the aforementioned points in the paragraph highlight the need for a well considered data protection and location strategy.
What Considerations Should You Make?
- Latency between the customer and the compute or latency between the application and the data. In order to manage against latency risk you need to have a well understood environment, combined with a strategy for how your applications will be migrated to public cloud, moved to colocation or not migrated.
- Data in the cloud can offer significant performance advantages and in some cases operational advantages over an alternative, especially assuming you’re running your applications where your data sits. However, the largest cost area for cloud use is often associated with data use (Storing, egressing, etc). Using a CSP for your data needs also means you are scrapping all but the most painful of exit strategies. Mitigating the risk of cost with a benefits analysis of latency and staffing models should always be a consideration.
- Data Resiliency (loss, availability or corruption prevention) is as good as the tactics used when data is captured. In most every case when it comes to critical data you’ll be creating a duplicate and a recoverable (real time or other) backup. Data replication between cloud and on premises potentially adds additional value in both providing protection and reducing latency for apps. In other words, even if you decide that your cloud based applications require the data to be local, having the data on premises offers additional benefits.
- I’m not one to question the capabilities or the enthusiasm of CSPs and their processes for protecting my data, I have seen that they largely provide excellent services. That being said, compliance belongs to the enterprise, not to the CSP. In some cases the only way to ensure you are compliant and can remain that way is if you have your data sets on premises, managed by you.
- Edge computing, with the exception of a “campus” based Edge means that no one place will be the only place you’ll need data. Developing applications, security, encryption and resiliency strategies will take on a whole new meaning. Where will your analytics be run and what will the performance expectations be? How distributed is the Edge for your business and your customers or partners? Developing a strategy that leverages a flexible, scaleable, independent solution might provide you the greatest ability to stay both protected and nimble. Most importantly, your solution choice must include common management tools that allow almost anyone in an organization to use in an emergency.
By no Means Comprehensive
This short little blog isn’t meant to be comprehensive, but rather to highlight what I believe are critical aspects of data management and governance (“Ownership” in Mark speak). As a business, access to your data where you want it, in the format you want it, with the protections and compliance necessary is minimum table stakes for business continuity. It’s also true that the ability of your businesses applications to be responsive to customers (employees or external customers) is more important to the bottom line than most people realize.
Don’t let anyone tell you that latency isn’t important. There will be cost effective opportunities to differentiate your services at the edge and those who do first, even if just the different between 60 & 20 ms in latency, will be the winners.