The world is cloud-obsessed.
Whether it’s streaming TV; or how you take, edit, store and share your photos; or your documents; or even how you communicate with your family and colleagues, the products you use will be sold as something to do with “cloud”. The reality though is that it’s all crap. The IT sector has a long history of (re)selling the same thing over and over under a different name, and “cloud” today is literally no different from software-as-a-service in the 1990s, or multi-tenant computing in the 1980s and earlier – it just has 21st-century marketing.
IT organizations surprisingly fall into this trap today, despite being insiders to the game. New projects and applications “must be cloud”, and every organization’s CIO is expected to have a “cloud strategy”. What that really means for the vast majority of them is that they want to use IaaS (infrastructure as a service), which is really just virtual machines with micro-billing, and that is a perfectly reasonable technical and cost objective.
Not all applications are loosely-coupled web-based services that can exist or behave consistently and predictably in a variable-latency, variable-performance, over-committed, shared compute environment; i.e. “the cloud” is not a fit-all model.
Furthermore, (public) IaaS is only cost-competitive if you don’t run it constantly. Compare like cloud instances against leased dedicated physical compute and you’ll find it costs more. Guess what? Most organizations fire-and-forget and end up running countless “permanent” IaaS instances – ironically exactly what the IT vendors want you to do to increase your overall spend.
“Cloud” may fit, and may play a part; but don’t be fooled into thinking it’s got to be the answer.
Unblue has decades of experience in designing and building suitable infrastructure whether it’s small web applications, or very large enterprise-grade, mission-critical, cross-region resilient applications.
Feel free to get in touch and discuss your requirements.