How Will Cloud Impact Your Existing Application Portfolio?
In a conversation with a vendor in the application portfolio management space the other day, we got on the subject of what "Cloud" means to them and by extension to their current and future customer base. My colleagues have written extensively on what it may mean to Oracle, Microsoft, IBM, HP and others, and our conversation began discussing the potential of cloud-computing from a vendor perspective – for instance:
- It opens a distribution channel for vendor products – rather than relying on a 6-month sales cycle, cloud provides an "infiltration" sales channel whereby a firm can load one or more applications and reap the benefits of application mining on a "pay as you go" basis. Good for the vendors, good for the clients – it's all good.
- How vendors fully make the transition to a pay-as-you-go model has some land-mines for vendors however. Over the time it takes them to transition from traditional models to a full pay-as-you-go model, vendors' balance sheets will lose the 6-figure product sales (revenues nose-dive), but gain more as infiltration sales demonstrate the value of the product, and use increases. This is a classic slow-dimes versus fast-nickels concept.
- There are tremendous indirect impacts to expense models as well. Eventually, vendors benefit by supporting only one (current) software version, or perhaps current and one back-release, and shouldn't have to perform extensive testing on dozens of combinations of release-levels on various operating system/database combinations. Suffice to say that all of the SaaS benefits accrue to vendors hosting software via the cloud.
The conversation turned from vendor benefits to the benefits that cloud may bring to all of you folks in IT shops in the public and private sectors – what impact will cloud have on the millions of applications that run your core business processes and transactions? There will eventually be an impact on your data centers – where your applications run – and those changes may impact your firms' attitudes toward staffing models – who maintains the applications and where? But those are indirect impacts to applications. What impacts do anticipate for your existing applications from cloud, and when do you think it will happen? Is it happening now? 1 to 3 years out? 3 to 5 years? For instance:
- Will you re-engineer them into services that can be deployed and consumed anywhere? Does that happen to all applications, or a select few, and if only a few – what criteria will you use?
-
Is cloud our modern-day Cretaceous-period-meteor that will snuff out an entire genre of monolithic applications like the mass-extinction of dinosaurs 65 million years ago?
(Yes, my tongue was implanted firmly in my cheek as I asked this question) We've been certain that would happen because of various technology advances in the past, and it has yet to happen. - Will cloud's impact to existing applications be more limited? Will it merely be accretive to the computing models currently in place, offering another option, but not wholesale replacement?
My goal for this blog entry is to encourage healthy debate on various aspects of cloud's impact to all those applications out there – the perceived benefits, risks, opportunities and threats. I'm asking you to offer your opinions and insights (or even just questions) about how cloud will impact the existing application portfolio. To get the ball rolling, I've seeded the blog entry with a few "for instance" points:
- Cloud will certainly affect packaged-application deployment options as it will affect many other kinds of software vendors. I can see that vendors who write their applications as consumable SOA services will be able to deploy their applications in a flexible manner. That will allow customers to custom-configure and use only the pieces they need.
- Custom-built apps, are a different animal entirely, methinks. While these points hold true for packaged apps as well, the devil is in the detail of integration points, data reliance, and perhaps up-stream and downstream workflow processes.
This should be enough to get the conversation started – I don't want my perceptions to dominate – what I really want to know is … "What do you think?"