Not surprisingly, your enterprise wants to embrace the public cloud. Also not surprisingly, you're not finding it trivial.
Oh, sure, writing apps to run on Amazon Web Services or Microsoft Azure is a snap if it's a new app that is "born" in the cloud. But what about the majority of your apps, those earthbound misfits that have happily (or not) lived in your data centers for years? What to do about them?
According to a new briefing from Forrester, while customer-facing apps tend to be simple to move to the public cloud, systems of record are almost always "old and either completely or substantially custom-built," which means "migrating these applications to public clouds requires substantial revision to achieve cloud's primary benefits."
Is the cost worth the bother?
To the clouds
To be clear, the question isn't whether enterprises will run their businesses in the public cloud. They will. As noted here on TechRepublic, public cloud adoption dramatically outpaces private cloud (another way of saying "enterprise data center") growth.
There are a variety of ways to measure that growth. The first, of course, is revenue. By that metric, AWS dwarfs the private cloud market all by itself, with private cloud pushing $7 billion while AWS now surpasses a $10 billion run-rate. If we look at VMs running in the cloud, as Gartner's Thomas Bittman has, the difference is even starker. According to Bittman, public cloud VMs have grown 20X over the years while private cloud growth sputters along at 3X growth.
SEE Public cloud crushing private cloud in growth and revenue (TechRepublic)
The reason is simple: "New stuff tends to go to the public cloud, while doing old stuff in new ways tends to go to private clouds." And there happens to be a lot more "new stuff," as enterprises look for ways to improve the customer experience.
Hence, as Forrester analysts John Rymer, Dave Bartoletti, and William Martorelli write in a new brief, The Cost Of Migrating An Enterprise Application To A Public Cloud Platform, "Evidence is mounting that public clouds will be the preferred— if not dominant—application platform for customer-obsessed organizations."
But that doesn't mean it will be easy.
At what cost?
How hard—or easy—it is to move an app to the public cloud largely depends on the kind of app it is, and how long it has been hanging around one's data center. As Forrester notes, "customer-facing apps for systems of engagement...typically employ lots of new code rather than migrating existing code to cloud platforms," making migration (if any) a straightforward affair. As a bonus, "the speed to market and responsiveness of these apps trump cost as the primary management concern."
So, for customer-facing apps, an enterprise generally needs to justify not having such apps live in the public cloud, rather than the reverse. But for so-called systems of record, or "conventional enterprise systems designed to contain the authoritative data source for a given piece of information," it's a very different story.
SEE Survey: Is the hybrid cloud living up to its potential? (TechRepublic)
As noted, Forrester said most of these apps "are old and either completely or substantially custom-built." Think manufacturing, CRM, ERP, financial, or other such applications. Forrester continues: "When a software-as-a-service (SaaS) replacement isn't an option, migrating these applications to public clouds requires substantial revision to achieve cloud's primary benefits—on-demand scaling, pay-for-what-you-use economics, global availability, and high security."
Such applications were built with very different priorities in mind, and developer productivity isn't one of them.
But it's those same developers that are going to bear the brunt of migrating such systems of record to the public cloud. By Forrester's estimate, the cost of public cloud is relatively small compared to the much larger cost of labor involved, which accounts for over 50% of total migration costs. As detailed in the brief, "[L]abor costs dwarf infrastructure and platform services costs in most of the migration projects we've reviewed."
The firm notes that these labor costs decline over subsequent projects as teams become familiar with the work involved, but make no mistake: The biggest cost (by far) of moving old apps to new public clouds is the people involved in effectuating the shift.
Given the cost of teaching old apps new tricks, it's a fair question as to whether it's worth it. Sure, doing so can result in lower ongoing operating costs, but will an enterprise ever recoup such costs?
Definitely maybe. Jonathan Feldman writes that "many legacy applications [can] leverage at least some of the technology advances inherent in infrastructure-as-a-service," thereby netting millions of dollars in savings while making applications much more agile.
There's also the matter of necessity: Yes, there are still "green screen" applications lurking in the bowels of enterprise IT within most large enterprises, but these aren't the systems that are keeping enterprises competitive. To keep up with the fast pace of today's business, public cloud is a requirement, not an option.
- Public cloud crushing private cloud in growth and revenue (TechRepublic)
- Everything you need to know about cloud in just one tweet (TechRepublic)
- Survey: Is the hybrid cloud living up to its potential? (TechRepublic)
- CIOs keep trying to defy cloud gravity (TechRepublic)
- How public cloud providers are making security a non-issue for app developers (TechRepublic)
Matt is currently head of the developer ecosystem at Adobe. The views expressed are his own, not those of his employer.
Matt Asay is a veteran technology columnist who has written for CNET, ReadWrite, and other tech media. Asay has also held a variety of executive roles with leading mobile and big data software companies.