Grain Silos
We get an avalanche of multicloud selling from storage vendors. We get promises and benefits of multicloud but from whose point of view?
Multicloud is multiple premises
This is an overly simplistic example how I created 3 copies of the same spreadsheet yesterday. I have a quotation on Google Sheets. A fairly complicated one. Someone wanted it in Excel format, but the format and the formulas were all messed up when I tried to download it as XLSX. What I had to do was to download the Google Sheets as ODS (OpenDocument Spreadsheet) format to my laptop, and then upload the LibreOffice file to my OneDrive account, and use Excel Online to open the ODS file and saved as XLSX. In one fell swoop, I have the same spreadsheet in Google Drive, my laptop and OneDrive. 3 copies in 3 different premises.
As we look to the behaviour of data creation and data acquisition, data sharing and data movement, the central repository is the gold image, the most relevant copy of the data. However, for business reasons, data has to be moved to where the applications are. It could be in cloud A or cloud B or cloud C or it could be on-premises. The processed output from cloud A is stored in cloud A, and likewise, cloud B in cloud B and so on.
To get the most significant and relevant copy, data from all premises must be consolidated, thus it has to be moved to a centralized data storage repository. But intercloud data movement is bogged down by egress fees, latency, data migration challenges (like formats and encoding), security, data clearance policies and many other hoops and hurdles.
With all these questions and concerns in mind, the big question mark is “Is multicloud really practical?” From a storage guy like me who loves a great data management story, “It is not. Multicloud creates storage silos“.
Storage Silos kills Data Unification
Storage vendors tout multicloud. Their enterprise storage business to on-premises data centers is dwindling and many have pivoted to the hybrid cloud business model. Hybrid is the new Black. The multicloud hype sprung up a few years back, because the storage vendors have little else to sell. Their alignment to one, or two or all of the Big 3 Clouds is a must to ensure their survival into the next 5 years or so. And multicloud is way for them to sell you their gear in either one of the clouds or on premises, as long as it is on their gear and their technology.
Behind the illusion of the touted benefits of multicloud, the challenges of unifying the data into a single, centralized copy can be unsurmountable when the storage of the data is spread across many premises. Storage silos.
We often hear of data is the new oil. Data is lifeblood. Data is the new currency that will drive digital transformation. However, we must also bear the responsibility that inaccurate, stale, irrelevant data, poorly managed data equates to poor decision making, and consequently, poor outcomes for the digital transformation initiatives. Storage silos.
People, process, technology
Thus we must look at the IT maxim of people, process and technology. Technology should always be last. We must understand that technology is fleeting. Technology comes and goes. The multicloud hype is fueling the technology part where organizations can be enamoured by promises of multicloud.
Organizations must see beyond the wools pulled over their eyes. The processes within the data managed in the organization and the storage repositories that store them are far more important. The people entrusted to keep the data safe and secure, relevant, useful must look to the reduction and even the elimination of storage silos, regardless of premises because data movement is costly. Data disunity and disaffiliation across storage silos in different premises is costly.
Stick with one
Hybrid clouds, which have been gaining acceptance in many organizations today, is about data storage in at least 2 premises, depending on workloads. One on-premises, and the other, in the cloud. Organizations must now extend their IT processes and policies in their enterprise data centers to work as transparently and as seamlessly with the cloud service provider of their choice. That practice can be easier said than done when it comes to data storage because data has gravity. Data has to be presented, shared, protected, preserved, archived in its data lifecycle and workflows within the organization, and every storage premises that the data lands on has a cost, a latency, a security access, a compliance and governance reliance and dependency with other pieces of data to its own.
In the midst of that extension, reliance, dependence and more, the challenges to create, maintain and manage a single, organization-wide data policy across a multi-premises, multicloud is not just practical.
So, in the face of a hybrid cloud strategy, I propose to stick with just one cloud storage provider to unify the data strategy and policies, and discard the storage silos sprouted in a multicloud environment.
Why use any clouds for storage? You can architect your central data repository in a multi-cloud adjacent location, with appropriate direct connection capability (to avoid egress $), and use the cloud for what it’s good at- elastic compute and platforms that can use that data, either in place or copied into that cloud for processing.
Multi-cloud isn’t about just about data. It’s about workloads that happen to USE data. Add in the growing need for edge architectures , and this problem grows exponentially. Stuffing your data in one cloud is going to prove to be an anti-pattern.