Whether apps that are graphic-intensive or require heavy-duty processing are suited to be accessed via cloud is less a matter of technical feasibility, and more an economic decision of forking out money to redo the network architecture.
The cloud model of software-as-a-service (SaaS) works "pretty well" for most applications, but performance tends to lag for apps requiring massive volumes of data or heavy-duty processing and graphic features, said Charles King, principal analyst at Pund-IT.
These two characteristics make such apps problematic, but not impossible, for cloud delivery. A huge data volume needs a robust wireless network performance, so using apps will be a "painful" experience in less-than-optimal mobile wireless environments, he explained.
As for "processing muscle", it is sometimes subjective, depending on certain types of tasks or users of an app. Some users might say they can do photo-editing on a tablet, "but I'd qualify such claims as hobbyist boasting rather than proof points that would be convincing to a business" such as a graphic design company, King noted.
Whether certain apps are not ideal for cloud delivery is actually a question of business, rather than technical, feasibility, according to John Brand, vice president and principal analyst at Forrester Research.
Latency as well as security and privacy concerns are the two biggest stumbling blocks to an optimal user experience of SaaS apps, regardless of function or features. Latency is dependent on whether or not the network architecture can capitalize on the benefits of cloud, without sacrificing end user experience; whereas security is a business risk issue, not a technical one, Brand explained.
"With the right underlying architecture, even low latency applications can be successfully moved to the cloud. If organizations find they have apps they can't move to the cloud, it's an economic decision based on redeveloping the architecture, rather than whether it is technically possible. At the end of the day, it's about choosing the right architecture that gets the job done for customers at the right price," he said.
In other words, it could cost the company too much to re-architect the entire solution for the cloud. The company may also be unable to make or reach a strategic long-term decision about which apps to move and why, he noted.
Besides economics, it could be a cultural decision as well, whereby the company or its customers fundamentally object to moving any apps to the cloud, he added.
The cloud is "remarkably adaptable", said Brand. "Whereas we originally thought business intelligence (BI) and analytics would probably never move to the cloud because of the large data volumes, we now see vendors providing exactly these kinds of managed services for various industries, and in some cases, these solutions were developed purely because of the cloud."
The same can be said for graphics-heavy games and rich media editing tools which have intensive processing requirements, he noted.
Steve Hodgkinson, IT research director, Asia-Pacific at Ovum, concurred: "It is more about the fit between business needs and the app, rather than the app itself. Pretty much any app can be delivered as SaaS if it is well-designed and well-operated and there is adequate Internet connectivity."
The degree to which an app is viewed as "pet-like or cattle-like" has a big influence on the economics of re-hosting it to a cloud environment, Hodgkinson pointed out. He referred to "pet apps" as those that are unique, highly-customized, and hard to replace, not least from being "tenderly" looked after by their developers. On the other hand, "cattle apps" are considered standardized and replaceable.
Analysts agreed that developers of apps seemingly ill-suited for SaaS delivery will eventually find some way to move apps to the cloud, for various reasons ranging from customer and market pressure to cost considerations.
Hodgkinson said: "There will still be many years ahead of us where enterprises prefer on-premise installed software because of legitimate needs or stick-in-the-mud reasons. And many vendors will still make good money out of selling traditional software to traditional buyers, but the problem is this will become a high-cost model for both vendors and buyers, and hence ultimately decline."
Being stuck with the higher cost structure with non-SaaS apps brings the burden of developing and testing software to run on a diverse range of hardware, and deal with app integration and interoperability challenges, the Ovum analyst explained.
Forrester Research's Brand noted that companies need to architect for the cloud "from the bottom up" in order to reap the full benefits. "That's why we're likely to see a "long tail" of on-premise applications for many decades to come. The question is not whether to go cloud or not. It's what to move to the cloud and when in order to get the full benefits and lower the overall risks. That's completely a commercial decision and will be different for almost every organization."
The rise in bandwidth and proliferation of multiple-device ownership is pushing companies such as Adobe to move more apps and services to devices and the cloud. For instance, Photoshop Touch is a tablet-optimized version of its same desktop software.
Adobe has also been shifting to cloud to offer its products. Its rich-media editing, such as Photoshop, and digital marketing software are offered as the Creative Cloud and Marketing Cloud respectively, on a monthly or yearly SaaS subscription.
Paul Burnett, Creative evangelist, Asia-Pacific at Adobe, said some of the apps included in the Creative Cloud, such as Photoshop, Illustrator, InDesign and Premiere Pro, require creating and editing large, rich media files. Hence these are downloaded to run locally on the user's computer, and users are notified whenever there are software updates. Other tools and services, including online storage, sharing and collaboration run in the cloud.
According to Burnett, this arrangement "provides users the best of both worlds". "The Creative Cloud enables them to access all the apps, tools and services, and also receive updates as they are available at no extra cost."
Asked how Adobe decides which and the way heavy-duty media-editing apps can feasibly be hosted on the cloud, Burnett replied: "We will continue to innovate in the cloud and devices space, building creative solutions for our customers when and where they make sense."