Executive Briefings

For Supply-Chain Apps, the Forecast Calls for More Cloud

Companies haven't come close to exploiting the potential of cloud technology for running critical software applications. But there are trade-offs to be considered, when deciding whether to trust your data and systems with an outside vendor.

Like so many "new" ideas, the cloud has been around for awhile.

The concept has had many names - software as a service (SaaS), hosting, virtualization. For some years now, companies have seen the merits of accessing at least some software programs outside their own servers and firewalls. Only recently, though, have they begun to take seriously the cloud as a secure source of applications to manage business processes across the supply chain.

Their reluctance until now has been understandable. Chief among their concerns is the issue of security. Businesses unfamiliar with remote hosting worry that sensitive data will be exposed on the network for all to see. They further wonder about the stability of off-site servers running critical processes (although the same concern can apply to computers on their premises).

Mike Dominy, research director with the Gartner consultancy, has been talking with clients about the cloud for nearly a decade. Some of that initial buzz was hype, given the drawbacks of internet technology and computing power back then. But some arose naturally from a desire for business-to-business integration and the flow of data between supply-chain partners, especially manufacturers and retailers. Dominy says the cloud became an "architectural representation" of the push to break down corporate walls and treat the multi-partner supply chain as a single, coherent entity.

The breakthrough, he says, came as virtualization technology gained momentum. The term refers to a departure from the traditional model of purchasing a server along with all of the accompanying software licenses for running business processes in-house. The problem lay in figuring out just how much capacity to buy. Companies would have to forecast future volume loads; typically they would acquire more than they needed on a day-to-day basis in order to ensure sufficient computing power during peak periods. The problem was especially acute in retail, where businesses face major scalability issues during Christmas and other seasons of heavy consumer activity.

Says Dominy: "As technology vendors and companies look continually to get more money out of their computing investments, there's a natural progress to move to the cloud." What's more, buyers can pay for applications according to their actual level of usage, without having to worry about the headaches associated with installing periodic updates.

The idea didn't mature overnight. What was needed was a flexible computing infrastructure - one that could be shared among multiple users in order to minimize idle capacity. By definition, says Dominy, cloud technology is "elastic and scalable." That's a tough goal to achieve when hardware and software are isolated within individual companies, no matter how large they might be.

The Cloud Takes Hold

As technology advanced to make the sharing of resources possible, software vendors began to score successes with hosted applications in certain targeted areas. One was customer relationship management, marked by the emergence of Salesforce.com. Another was transportation-management software, where SaaS has been an option for nearly 10 years, says Dominy. Today, there's scarcely a supply-chain application that isn't being offered in the cloud, although the acceptability of that option varies according to the nature of the specific tool.

One obstacle to wider acceptance of cloud technology has been the fuzziness associated with the term. Some vendors have used it interchangeably with older labels such as SaaS and hosting, while others argue that "cloud" indicates a deeper sharing of resources by multiple partners. Dominy says "hosting" involves the outside maintenance of infrastructure that is running applications for a dedicated customer. Such an arrangement might relieve the end-user of the burden of maintaining servers on site, but it doesn't take full advantage of the flexibility and efficiency of true cloud technology.

The nature of the cloud-delivery model creates challenges in some areas of the supply chain, Dominy says. One is the execution-oriented space, particularly manufacturing and warehouse management. In the case of the latter, companies can't tolerate any delay of data between an operator and a piece of machinery. While some WMS vendors claim this isn't a problem with cloud-based applications, "if you have a high-volume distribution center, with many dozens of operators running around doing [radio frequency] scans, the potential for breakdown and latency in the communications flow ... is a risk," says Dominy.

The same problem goes for the manufacturing environment, where automated machinery and process controllers require the instantaneous exchange of data. But even if there is no communications lag, the larger question is whether a cloud-delivery model makes sense for all types of users. Reliance on the cloud can reduce information-technology costs, Dominy says, but that's not necessarily the number-one priority of supply-chain executives. They're more concerned with end-to-end visibility and gaining better control over product, data and cash across the network. To the extent a cloud-based system achieves that goal, a company will embrace it.

Another confusing aspect of the debate over cloud technology is the failure to distinguish between public and private clouds. A public cloud involves the maintenance of all infrastructure, applications and data outside the user's facilities, with multiple parties sharing that capacity (albeit with strict security measures in place to prevent the unwanted exposure of proprietary data). In a private cloud, says Dominy, the applications are running on a set of resources that is dedicated to a single customer. While that option addresses the concerns that a company might have about data security, it doesn't deliver the efficiencies and scalability of a public cloud.

The ERP Debate

Of all types of software that can be offered in the cloud, enterprise resource planning probably generates the biggest questions among potential buyers. ERP applications form the backbone of many companies' financial systems; by definition they involve highly sensitive information. In addition, users worry about the economics of subscription-based pricing over lengthy periods.

For now, says Dominy, companies are more likely to accept the notion of ERP being delivered in a private cloud, with its extra level of protection. As for the desirability of a one-time purchase versus open-ended subscription, "there is absolutely a break point when it becomes more expensive to procure it in a subscription-based model," he says. "It has to do with how the pricing model is set up by the vendor, and how much volume you're talking about." Subscription tends to be the cheaper option in small companies with lower volumes, while businesses with large infrastructures need to examine their total cost of ownership over a three- to five-year horizon. Such a calculation might well advise against the acquisition of ERP in the cloud.

The scariest aspect of cloud delivery is the feeling among end users, justified or not, of a lack of control over their own applications. Many were shaken by the failure earlier this year of a major data center operated by Amazon.com on behalf of multiple clients. The outage left some companies' operations paralyzed for more than a day.

Despite the headlines generated by the failure, Dominy doesn't believe that the incident was a major setback to the advancement of cloud technology. "It just highlights that there are issues and risks associated with it that companies need to account for." Some users of cloud-based applications are protecting themselves by paying for redundant storage in more than one set of servers. Others have manual processes standing by in the event of a temporary outage. All companies need a backup plan, says Dominy, including those with critical applications running on their own servers.

For the future, he sees companies embracing multiple models of delivery for their applications, including the use of both public and private clouds. Global trade management is one area that's ripe for the cloud, he says, and transportation-based systems will continue to migrate to that option. "I think you'll see a mixed environment even five to 10 years out," Dominy says.

What he doesn't see anytime soon is the easy availability of all supply-chain applications on mobile devices. "The majority of companies we talk with say it's possible," Dominy says, "but it's going to be a real stretch with massive investment, to get to the level of a cloud-based delivery model on your iPad or iPhone."

Resource Link:
Gartner

Like so many "new" ideas, the cloud has been around for awhile.

The concept has had many names - software as a service (SaaS), hosting, virtualization. For some years now, companies have seen the merits of accessing at least some software programs outside their own servers and firewalls. Only recently, though, have they begun to take seriously the cloud as a secure source of applications to manage business processes across the supply chain.

Their reluctance until now has been understandable. Chief among their concerns is the issue of security. Businesses unfamiliar with remote hosting worry that sensitive data will be exposed on the network for all to see. They further wonder about the stability of off-site servers running critical processes (although the same concern can apply to computers on their premises).

Mike Dominy, research director with the Gartner consultancy, has been talking with clients about the cloud for nearly a decade. Some of that initial buzz was hype, given the drawbacks of internet technology and computing power back then. But some arose naturally from a desire for business-to-business integration and the flow of data between supply-chain partners, especially manufacturers and retailers. Dominy says the cloud became an "architectural representation" of the push to break down corporate walls and treat the multi-partner supply chain as a single, coherent entity.

The breakthrough, he says, came as virtualization technology gained momentum. The term refers to a departure from the traditional model of purchasing a server along with all of the accompanying software licenses for running business processes in-house. The problem lay in figuring out just how much capacity to buy. Companies would have to forecast future volume loads; typically they would acquire more than they needed on a day-to-day basis in order to ensure sufficient computing power during peak periods. The problem was especially acute in retail, where businesses face major scalability issues during Christmas and other seasons of heavy consumer activity.

Says Dominy: "As technology vendors and companies look continually to get more money out of their computing investments, there's a natural progress to move to the cloud." What's more, buyers can pay for applications according to their actual level of usage, without having to worry about the headaches associated with installing periodic updates.

The idea didn't mature overnight. What was needed was a flexible computing infrastructure - one that could be shared among multiple users in order to minimize idle capacity. By definition, says Dominy, cloud technology is "elastic and scalable." That's a tough goal to achieve when hardware and software are isolated within individual companies, no matter how large they might be.

The Cloud Takes Hold

As technology advanced to make the sharing of resources possible, software vendors began to score successes with hosted applications in certain targeted areas. One was customer relationship management, marked by the emergence of Salesforce.com. Another was transportation-management software, where SaaS has been an option for nearly 10 years, says Dominy. Today, there's scarcely a supply-chain application that isn't being offered in the cloud, although the acceptability of that option varies according to the nature of the specific tool.

One obstacle to wider acceptance of cloud technology has been the fuzziness associated with the term. Some vendors have used it interchangeably with older labels such as SaaS and hosting, while others argue that "cloud" indicates a deeper sharing of resources by multiple partners. Dominy says "hosting" involves the outside maintenance of infrastructure that is running applications for a dedicated customer. Such an arrangement might relieve the end-user of the burden of maintaining servers on site, but it doesn't take full advantage of the flexibility and efficiency of true cloud technology.

The nature of the cloud-delivery model creates challenges in some areas of the supply chain, Dominy says. One is the execution-oriented space, particularly manufacturing and warehouse management. In the case of the latter, companies can't tolerate any delay of data between an operator and a piece of machinery. While some WMS vendors claim this isn't a problem with cloud-based applications, "if you have a high-volume distribution center, with many dozens of operators running around doing [radio frequency] scans, the potential for breakdown and latency in the communications flow ... is a risk," says Dominy.

The same problem goes for the manufacturing environment, where automated machinery and process controllers require the instantaneous exchange of data. But even if there is no communications lag, the larger question is whether a cloud-delivery model makes sense for all types of users. Reliance on the cloud can reduce information-technology costs, Dominy says, but that's not necessarily the number-one priority of supply-chain executives. They're more concerned with end-to-end visibility and gaining better control over product, data and cash across the network. To the extent a cloud-based system achieves that goal, a company will embrace it.

Another confusing aspect of the debate over cloud technology is the failure to distinguish between public and private clouds. A public cloud involves the maintenance of all infrastructure, applications and data outside the user's facilities, with multiple parties sharing that capacity (albeit with strict security measures in place to prevent the unwanted exposure of proprietary data). In a private cloud, says Dominy, the applications are running on a set of resources that is dedicated to a single customer. While that option addresses the concerns that a company might have about data security, it doesn't deliver the efficiencies and scalability of a public cloud.

The ERP Debate

Of all types of software that can be offered in the cloud, enterprise resource planning probably generates the biggest questions among potential buyers. ERP applications form the backbone of many companies' financial systems; by definition they involve highly sensitive information. In addition, users worry about the economics of subscription-based pricing over lengthy periods.

For now, says Dominy, companies are more likely to accept the notion of ERP being delivered in a private cloud, with its extra level of protection. As for the desirability of a one-time purchase versus open-ended subscription, "there is absolutely a break point when it becomes more expensive to procure it in a subscription-based model," he says. "It has to do with how the pricing model is set up by the vendor, and how much volume you're talking about." Subscription tends to be the cheaper option in small companies with lower volumes, while businesses with large infrastructures need to examine their total cost of ownership over a three- to five-year horizon. Such a calculation might well advise against the acquisition of ERP in the cloud.

The scariest aspect of cloud delivery is the feeling among end users, justified or not, of a lack of control over their own applications. Many were shaken by the failure earlier this year of a major data center operated by Amazon.com on behalf of multiple clients. The outage left some companies' operations paralyzed for more than a day.

Despite the headlines generated by the failure, Dominy doesn't believe that the incident was a major setback to the advancement of cloud technology. "It just highlights that there are issues and risks associated with it that companies need to account for." Some users of cloud-based applications are protecting themselves by paying for redundant storage in more than one set of servers. Others have manual processes standing by in the event of a temporary outage. All companies need a backup plan, says Dominy, including those with critical applications running on their own servers.

For the future, he sees companies embracing multiple models of delivery for their applications, including the use of both public and private clouds. Global trade management is one area that's ripe for the cloud, he says, and transportation-based systems will continue to migrate to that option. "I think you'll see a mixed environment even five to 10 years out," Dominy says.

What he doesn't see anytime soon is the easy availability of all supply-chain applications on mobile devices. "The majority of companies we talk with say it's possible," Dominy says, "but it's going to be a real stretch with massive investment, to get to the level of a cloud-based delivery model on your iPad or iPhone."

Resource Link:
Gartner