Last week, Cloud Commons hosted a Twitter Chat on the end of Cloud Computing. If you’re not familiar with a tweetchat, they are discussions hosted on Twitter where people can join at a specific time by following a certain hashtag. The Cloud Commons tweetchats usually have around ten panelists and have been kicked off with a few thought-provoking questions. The participants then respond and share ideas in real time. The discussion is focused enough to be useful – 1 hour session, responses limited to 140 characters, but large enough to capture different perspectives.
This week’s tweetchat began with several questions:
- Adoption rates are rising for private cloud. Is this a stepping stone to hybrid/public cloud?
- What needs to happen before enterprises start to fully embrace cloud computing?
- What does the future model for enterprise cloud adoption look like?
- What should CSPs be doing more of to meet the needs of the enterprise?
- What needs to happen so that cloud becomes so ubiquitous that it’ll no longer be referred to as cloud? When will it happen?
The first question, “Is private cloud a stepping stone to hybrid/public cloud?” drew approximately 32 tweets. From the transcript, it appears as though participants in the marketplace are improving their understanding of cloud computing in terms of service and delivery models (private, public, hybrid, IaaS, PaaS, SaaS). The popular viewpoint was that private cloud is not exactly a stepping stone to hybrid/public cloud. A few tweets took the position that private cloud is seen as an alternate path to hybrid/public cloud. Many tweets indicated that IT departments want to retain tight control of their environment. Interesting tweet… “private cloud does not necessarily mean on-premises.” More on this later.
47 tweets in response to the second question, “What needs to happen before enterprises start to fully embrace cloud computing?” Overwhelmingly, the responses in this part of the chat were filled with terms like “services led,” “business value,” “SLA,” and “reduce FUD.” The responses to question 1 covered some territory here as well – enterprises will fully embrace cloud computing if and when they agree to give up some control of their infrastructure. There was an interesting tweet that mentioned transparency – “…it’s not always about control, as it is transparency.” We would argue that transparency is not needed here. To me, full transparency would require that the business is able to access minute detail about infrastructure, such as the amount of RAM installed on the application server that runs their slice of CRM at Salesforce.com. The business should be hidden from this kind of detail. Abstraction plays heavily here. So, we don’t need transparency as much as we need subtraction. What is an important concept that provides abstraction? You guessed it, Service Level Management. The GreenPages view is that processes need to improve before enterprises start to fully embrace cloud computing. See my earlier post, “What Should I Do about Cloud?” that goes in to much more detail on this topic.
I count about the same number of tweets in response to question 3 as I do question 2. Question 3 was a little more open-ended, so a critical mass of ideas never really took shape. The GreenPages’ view is that cloud computing will evolve to look like modern supply chains that can be seen in other industries, such as manufacturing. Enterprises may purchase IT Services from a SaaS provider, Salesforce.com for example. Salesforce.com may purchase its platform from another PaaS provider. That PaaS provider may purchase its basic infrastructure from an IaaS provider. Some value is added at each level, as the IaaS provider becomes more experienced in providing only infrastructure. The PaaS provider has an extremely robust platform for providing only a platform. The SaaS provider may ultimately become an expert at assembling and marketing these components into a service that provides value for the enterprise that ultimately consumes it. Compare this to the supply chain that auto manufacturers leverage to assemble a vehicle. In the early days of manufacturing, some companies produced every part of a vehicle, and assembled it into a finished product. I can think of one prominent example where the work to assemble a finished automobile took place in a single factory around the River Rouge in Detroit. Fast forward to present day, and you’ll be hard pressed to find an auto manufacturer who produces their own windshield glass. Or brake pads. Or smelts their own aluminum. The supply chain has specialized. Auto manufacturers design, assemble, and market finished vehicles. That’s about it. Cloud computing could bring the same specialization to IT.
Most tweets in response to question 4 were clearly around Service Level Management and SLAs, mitigating unknowns in security, and avoiding vendor lock-in. We agree, and think that a standard will emerge to define IT services in a single, consistent format. Kind of like OVF, the Open Virtual Machine Format, for virtualization. I can see an extension to OVF that defines a service’s uptime requirements, maximum ping time to a database server, etc. Such a standard would promote portability of IT Services.
Question 5 really went back to the topics discussed in question 3. When will enterprises embrace cloud? When will cloud computing become ubiquitous?
Right now, Corporate IT and The Business are two individuals living in a virtual “company town.” What I mean is that customers, (the business) are forced to purchase their services from the company store (corporate IT). GreenPages’ view is that there is a market for IT services and that emergence of cloud computing will serve to broaden this market. We recommend that organizations understand the value and costs of providing their own IT services in order to participate in the market – just like the business does. Overall, another insightful chat with some intelligent people!