Enterprise Cloud Migration: What if We’re All Wrong? | Gaining Altitude in the Cloud

Posted On October 6, 2011

This blog originally appeared on Sandhill.com. Read the original post.


Current conventional wisdom suggests that enterprise adoption of cloud services will accelerate as service providers and offerings become more “mature” and “enterprise friendly.” Adoption will grow and extend beyond initial test/dev website, and backup use cases as enterprises become more comfortable with cloud services. The common belief is that, over time, cloud will in fact become a strategic component of most IT environments but that it will be a decade-long (if not longer) transition. Most also believe that data security, privacy, and audit issues significantly constrain some verticals such as healthcare and financial services from effectively migrating in the near term, particularly to cloud platform and infrastructure services.

But as we discover far too frequently, conventional wisdom often turns out to be quite wrong. While adoption rates for new technologies tend to be overestimated in the short term and underestimated in the long term, it’s an interesting exercise to think about the factors and unexpected developments that could dramatically accelerate enterprise migration to the public cloud.

Let’s consider some of the basic assumptions many in the market make around enterprise and the cloud.

What if enterprises architect around SLAs?

The terms of many current cloud service provider SLAs are effectively meaningless. The burden of proof often falls on the user to fully document service interruptions and outages. Even if proven, compensation for violations often equates to a slap on the wrist at best. But what if enterprises come to the conclusion that SLAs are the wrong way to think about ensuring availability?

The highly visible Amazon outage in its Northern Virginia data center in April resulted in significant interruption and service degradation for users of Quora and FourSquare, while other websites and companies appeared to suffer no impact. The reason? Availability through redundancy. Rather than relying on SLAs, many unaffected companies simply architected redundancy through failover approaches that rolled to other Amazon data centers or service providers. What if enterprises decide that pushing cloud service providers on SLAs is akin to beating a dead horse and, instead, simply decide to take the SLAs as a given and architect around them?

What if enterprises standardize to conform to cloud service provider offerings?

Enterprises historically have been addicted to IT customization – both in what they buy, and how they buy it. Service providers that weren’t willing to modify offerings, pricing, or contract terms for large enterprise buyers were quickly shown the door. Many believe that enterprises will never migrate to cloud services that are essentially “take it or leave it” propositions to the customer.

Yet in many cases, enterprises have driven customization in processes, applications, and services that in fact add little or no business value. Cloud is opening many enterprises’ eyes to the fact that there may in fact be significant value in using cloud services as a lever to drive standardization across the organization, particularly for non-strategic applications and processes.

What if enterprises learn to live with standardization and limited configuration, and dramatically streamline support for non-strategic applications and assets?

What if data security and privacy issues are mitigated?

Data residency, security, and privacy issues are providing significant cloud migration constraints for some global enterprises, particularly those in compliance-sensitive verticals like healthcare and financial services. But what if these barriers were significantly reduced or fully eliminated?

Salesforce.com recently gave a glimpse into one way this may happen through the recent announcement of its Data Residency Option (DRO), which gives customers the ability to keep data on-premise behind their firewall while providing encrypted access to the Salesforce.com cloud application. Some cloud infrastructure service providers, like Savvis and Rackspace, offer dedicated hosting and private/public cloud services in the same data center, enabling hybrid models that support data “ownership” and the benefits of dynamic bursting into public cloud models.

While enterprise customers are seeking more transparency, Amazon has in fact achieved compliance with FISMA, HIPAA and PCI DSS and other standards. Some in the audit community are also discussing the need to reexamine common policies and controls in light of cloud services and architectures. The net net? Data security, privacy, and residency issues may end up being addressed faster than expected. What would happen to adoption if these concerns were taken off the table?

What if pricing for common IaaS services drops by 50 percent?

To date, cloud service providers have very effectively used private cloud economics as a pricing umbrella for their public cloud services. The result? Highly attractive margins for current cloud providers and an onrush of new providers. If microeconomics holds here (and I don’t know why it wouldn’t), pricing for public cloud infrastructure services will begin to drop, and potentially dramatically. Amazon has already established a pattern of driving consistent reductions in pricing for its core cloud services. What will happen when new entrants get aggressive in trying to grab share? The enterprise business case ROI for cloud service migration could be much more compelling in the very near future.

What if mission-critical applications migrate first?

The assumption is that adoption of cloud applications starts at the edge with line-of-business and functional applications and then, over time, migrates to more strategic and mission-critical applications. But what if CIOs determine to go in the opposite direction?

Examples exist of large global enterprises that have migrated to cloud service providers that offer hosted private cloud SAP ERP services in conjunction with community cloud spiking environments. While the common belief is that mission-critical apps will be the final frontier for enterprise cloud migration, what if it turns out to be the first?

All of these scenarios are unlikely to play out as described, but we can be sure that the conventional wisdom will be wrong in a market that is evolving as rapidly as enterprise cloud. Current expectations for the rate and pace of adoption are based largely on past trends in enterprise technology, which is probably a bad assumption in itself. It is increasingly clear that adoption curves for new enterprise technologies are actually accelerating.

I bet my money that the pace of enterprise cloud adoption will surprise many … it will be interesting to see what unexpected scenarios might open up the floodgates.

Everest Group Executive Viewpoints icon Related Articles