Yesterday I met two US government clients (one state and one county) and we discussed – amongst other things – the promise and potential of cloud computing. As I usually do, I asked them what they had in mind and what candidate applications were on their radar screen for possible deployment onto the cloud.
Not surprisingly their requirements and the characteristics of cloud computing (scalability, elasticity, metered by use, shared, accessible through Internet technologies) did not fully match.
Equally unsurprisingly, they were very interested in having my take about the Gmail deal in Los Angeles (see this interesting article on Government Technology). I told them that I did not speak to either LA officials nor to Google, but that I was quite surprised with some of the terms of the agreement reported by the press, such as the detail on unlimited and liquidated damages in case of a breach of Google’s servers.
It looks like government folks are looking at cloud as something that is unrelated to other ways of sourcing their infrastructure and application needs. On the one hand there are legacy systems, mission critical workloads, existing contractual arrangements with vendors that build and/or manage and/or operate technology solutions, which is the “traditional” world of IT. On the other hand there is the “cloud”, often with some of the same vendors, but with a completely different set of issues and decision-making approaches. Almost as if government organizations were forced to comply with an abstract requirement of using the cloud. Whether this comes from strategic directions at the federal level, or the excitement of an elected official in a state or county, or an unbearable budget pressure or urgency, cloud deployment discussions seem to take place in a sort of different dimension, where some obstacles and risks – such as security and reliability – as well as benefits – such as cost savings –. are overemphasized or treated outside the specific organization context.
One question I get a lot these days is whether Gartner has data indicating cost savings for different categories of workloads. As usual, my reaction is that metrics are useless unless they are put in context: how much will be saved by moving to the cloud depends on a long list of factors, many of which relate to the organization’s readiness to procure and consume cloud-based services. So I would argue that for two different organizations moving to the cloud may lead to very different cost savings (or otherwise) as well as risks.
Almost every client conversation seems to indicate the need for guidance about whether and how to select cloud-based solutions versus different sourcing models. Whether you name it a readiness assessment, selection criteria or portfolio segmentation, that’s where most are banging their heads on. Some vendors have proposed frameworks, but my sense is that they are cloud-centric, pretty much like their clients’ conversations.
Clouds do not exist in isolation, but in a continuum of models that range from DIY data centers and apps to cloud-based services, going through a variety of other, more traditional models. Understanding the nuances in this continuum, and which approach is most appropriate for which application need is key to both better understand where cloud services can create value, and articulate a roadmap to move toward increasing adoption of cloud services.
So the right question to ask is not “how do I use the cloud”?” but “what exactly do I need it for?”.
Category: cloud Tags: