Choosing between public and private cloud calls for solid financial analysis, taking into account a variety of important factors.
Enterprise datacenters have evolved tremendously over the years. They began as rooms primarily designed to house the mainframe before transitioning to smaller, distributed servers. Next came virtualization, which allowed for a consolidation of servers on shared hardware resources. For many enterprise IT groups, the first step into virtual environments was really a re-creation of the distributed model using many virtual servers rather than dedicated hardware.
In the middle of this transition to virtual environments, the cloud evolved. I say "evolved" because hosted services offered via the Internet have been around for a long time. However, service providers have reduced costs, added significant capabilities to their offerings, and leveraged virtualization to create products such as IaaS, PaaS, and SaaS. These cloud solutions are all the buzz, right now, because they have the potential to significantly reduce costs, improve the efficiency of IT departments, and be highly profitable for service providers.
Enterprise IT departments are starting to take notice and learn from these public clouds. Rather than simply duplicate the distributed environment on virtual servers, they are now using the power of the virtual environment's resource schedulers and other features to deploy private clouds. Instead of manually provisioning a new server for every need, they are implementing tools that enable internal clients to easily provision or decommission needed services on top of existing servers, and ensure that hardware resources are allocated or freed up as needed to accommodate these services.
So which is better -- public or private cloud? The simple answer from Microeconomics 101 is whichever one is economically more efficient. Gauging efficiency, though, means considering not just cash outlay, but also factors such as administration costs, maintenance, uptime, performance, security, and disaster recovery. Public clouds will prove more efficient for some organizations (and some applications) and private clouds more efficient for others. A good way to appreciate this is to consider an analogy based in logistics.
Consider a company called Ship-it, specializing in logistics and very good at it, too. By utilizing Ship-it's services, many retailers are able to eliminate completely the need to acquire expensive warehouses and trucking fleets and to hire employees to manage the flow of products from suppliers to far-flung retail stores. These retailers are able to benefit from Ship-it's economies of scale and pay for only the portion of that firm's warehouses and trucks that they use. For many companies, in other words, it is more efficient to use Ship-it's services than trying to duplicate those logistics operations in-house.
Megamart, meanwhile, is in the business of retail sales. It is not a logistics firm, but its superior skill in logistics is much of the reason that it is able to keep prices low and compete well in the marketplace. Were Megamart to use Ship-it's services rather than handle logistics itself, the retailer would likely see merchandise take longer to show up on its retail shelves, and this delay would effectively increase its costs, reduce cashflow, and depress margins.
As in this logistics example, which cloud solution is most efficient for a particular IT department will likely depend on that department's current capabilities and maturity. For one company, it may be best to move as many services as possible to the public cloud. At another enterprise, where the IT department is at a high maturity level, implementing a private, in-house cloud may prove more efficient. And many companies, of course, will likely benefit from a hybrid solution, for even the company with a strong IT department may find areas where using public cloud services enables it to free up resources that then can be devoted to more-profitable activities.
Ultimately, IT groups need to be open to both solutions, public and private. They should perform a proper, full analysis, and use the solution that looks most efficient for their business.
By the way, we had a good chat about this topic and the transcript of that session is archived online for anyone interested in reading it. It's right here.
Yes, that seems like the key item, a business plan that takes into consideration all factors and costs and benefits. Moving to the cloud is a big decision.
I suppose, though, there is a question here which we've not necessarily touched on: Is it always that case one or the other, public or private cloud, will be obviously better than the other? Is it inevitable that one will clearly win out over the other? If not, what should one do, how does one choose? What would be the threshold?
I believe that the implications are different, but they will both create a change and have some impact. Some services will transition easier than others, which is why it is important in any case to write a business case and project plan.
My guess is that choosing between a private and public infrastructure cloud is a different proposition than choosing, say, platform-as-a-service. The former is a low-level entity while the latter has major implications for developing and deploying apps.
It sounds like good advice, to try out the cloud with a well-chosen trial app or two. That way, you can better understand the economics and workings of the cloud services and not take the supplier's word for it.
From all I can tell, I think we will be seeing more storage, especially the longer-term variety, move to the cloud.
Of course, that doesn't solve the problem of where to run apps, where to keep servers, etc.
I certainly don't envy today's IT manager, with such momentous decisions to make. On ther other hand, there's a good deal of money to be saved for many companies. Look at Netflix, for instance, which is primarily in the cloud. I guess you could think of that firm as dependent on logistics, but it actually rents the trucks and warehouses (up in the cloud, at Amazon and some content delivery nets) and manages them itself, vs. outsourcing the whole thing.
I agree that it may not be completely black and white. There are a lot of moving parts to the equation that can be difficult to lockdown, but with some effort, a team can make the appropriate decision for their company. The key is to eliminate "excuses" why one is better than the other, and look at the real factors after performing some upfront research.
Putting a single service in the public cloud to better understand how things work and to gain experience is a good idea. However, the team must put their full effort into the project, and they should use an externally-facing service that is well suited for this environment. This type of strategy will help to maximize their success, and also provide for a strong learning environment.
Once the team moves through this process, they may be surprised what they find. At one company with a strong IT team, they found that it was more efficient for them to throw their archival storage into the cloud. They were required by government regulations to maintain certain data for up to 30 years, and they found that they were expending too much time dealing with backups, mirrors, and maintenance of this environment. While this is not a service typically thought of as being suited for the cloud, they found that moving it allowed them to focus on more critical items for the company.
INteresting article, Matthew. I wonder if you have any thoughts on how users should think about one kind of cloud being an on-ramp for the other. It seems to me that a company might try out cloud computing at, say, Amazon and once it had determined that this had some merit, it might want to bring some of that capability in house. Or vice-versa. In other words, perhaps the decision is not black and white, one or the other but instead, a mix.
Also, I gather that up to a certain level of capacity, renting in the public cloud is a good choice, but then it becomes less expensive to run one's own cloud. But then again, Netflix is pretty much all in the cloud, and it runs a big set of infrastructure, while Facebook is not. Who knows?
Making the calculation even more complicated is the fact that cloud computing is fairly new, and expertise with it is scarce, and the technologies involved are all evolving failry quickly. It's moving target.
Unlike most professions, the IT industry hasn't settled on a few all-important certifications for its practitioners. Mainly, vendors run the show. But this is changing, and that's a good thing.
To save this item to your list of favorite IT Software Community content so you can find it later in your Profile page, click the "Save It" button next to the item.
If you found this interesting or useful, please use the links to the services below to share it with other readers. You will need a free account with each service to share an item via that service.
Tweet This
