I recently attended a seminar put on by a consortium of hardware vendors extolling the values of private cloud architecture. They were introduced by a reputable industry analyst group that had done surveys of businesses in regards to their adoption of private cloud technology. Not surprisingly, the analysts and the vendors talked about how many organizations are looking to build a private cloud environment of their own. After listening to the pitch, I thought I would compare the two methodologies.
Just so I can lay all my cards on the table upfront, I have spent the last 15 years as a presales architect in the business of hardware sales. I have recently been working with NTT America on their public cloud services, so I think I have a pretty good look at both private and public cloud options. My personal opinion is that most technology will end up in a public cloud (as SaaS, PaaS or IaaS). Having said that however, I think there is room for discussion.
For the most part, the vendors that put on the seminar were probably right about businesses taking their first cloud steps toward their own private clouds. The reasons for this are pretty obvious. Many organizations already have at least a portion of the necessary infrastructure in place and have done some virtualization. So the next logical step would seem to be making their current infrastructure into a private cloud. There may also be some political motives for maintaining the status quo and extending the existing infrastructure that they own.
Fear and Politics
- Jobs – People get nervous about jobs when they start thinking about their hardware going somewhere else. There is an entire group of people in most organizations that do the systems administrative work, as well as the cabling and racking and stacking. There is also another group that manages the physical infrastructure and who would be affected by a movement to a public cloud provider.
- Loss of control – For many years I have worked in datacenters. There is nothing like the happiness you feel when you see the brightly flashing lights and hear the sound of the spinning disks and feel the warmth from the back of the servers. It is like a security blanket. And if I sound like I am being sarcastic, I am. Being inside of a datacenter is a pain in the butt, but having it close by makes people feel better. They feel that if something were to go wrong, they could run right into the datacenter and fix it. For example, I was recently working with a customer who was looking to create a virtual desktop environment. When we went through some of the basic sizing questions we discovered that the environment would never scale beyond 50 to 100 users. If you design a VDI solution for full redundancy, you are typically looking at a minimum of 3-5 servers and backend storage. This was a small company and looking at the numbers I didn’t think it made sense to build something out. When I suggested that a more appropriate solution might be to look at a cloud offering for building a VDI environment, the customer wouldn’t entertain the idea because of the control factor.
- Security – If it is behind my firewall, then it is protected. If you go back to my staffing argument from above, there are whole teams of security engineers working to make each corporation safer. They also have a fear of job displacement.
- Hardware vendors have less of an incentive to push public clouds because it diminishes the amount of people they can sell to. If all computing were to be consolidated to a handful of vendors, then in essence the hardware teams that sell to SMB and other businesses would be out of work or chasing the big guys. So for the most part it is not in their best interest to push a public cloud model.
So fear is what keeps organizations creating their own private clouds—fear of losing jobs, losing control, security breaches, etc. I would assume that much of the fear is based on the newness of cloud technologies and may not be completely warranted. If a survey shows many companies are looking to create a private cloud, it could be inaccurate. Asking whether or not a customer is looking at a private cloud solution may be leading. The question should be, “what is the better option for the customer?”
Public Clouds offer users an alternative to the traditional way of doing things. There are many advantages to using a public cloud over building your own. However, based on some of my previous comments, it may not seem like an obvious choice for some organizations. For example, if you have a datacenter that already has lots of growth built-in to the design and you have done your best to consolidate your infrastructure, the public cloud may seem too risky.
There are some benefits you do get from using a public cloud that you wouldn’t get if you built your own:
- Infrastructure – You no longer have to be concerned with managing and monitoring the infrastructure. So if a new series of servers comes out with different types of power supplies or requires more wattage, it is no longer your problem. You also do not have to worry about maintaining the space from a facilities perspective. No diesel storage, no backup generation, no running of fiber into the building so you can have high-speed connections. That is all taken care of by your vendor.
- Migrations – You now no longer need to worry about the cost to upgrade your environment. Every 3 or so years, most organizations refresh their equipment. That means that there is a capital outlay every few years. And during that timeframe, organizations without the public cloud need to have a group of people on staff to manage migrations.
- Jobs – It is true that with the public cloud, job qualifications will change for your organization. Fewer people will be needed to maintain the basic infrastructure. Those people could, however, be redeployed to provide higher value to your organization.
- Location independence – In my private cloud discussion I highlighted the enthusiasm I had for spending my day in the datacenter. Those days are a thing of the past. Remote access is now expected. By having the ability to spin up new machines and create fault tolerant environments in the cloud, I think this argument is moot. However, there are additional security features a business might want to include within the environment, so this is not a complete win for public clouds.
Public clouds have their downside as well.
- Virtualization type – If you have already invested in virtualization, there may be some additionally required steps to get your environment to the cloud. Some vendors base their solution off of different hypervisors, which may mean that a conversion is needed.
- Ability to migrate in and out of the environment – It takes time and expertise to move to and between cloud vendors. As mentioned above, there isn’t one standard for virtualization, so it may take more effort to get there. Generally, it is easier to start from scratch.
- Storage options – You will be using whatever the vendor decided to buy, so if you don’t like what you have, there isn’t any way to change it.
- Security – While this fear may not be founded in reality, sharing infrastructure with other companies does make people nervous
I could go on with a list of 100 different reasons why you would choose one technology over the other. I personally think that in the next 20 years, the ROI for public cloud providers will be too compelling to build your own private cloud. And why would you want to deal with the headache? On the other hand, while the public cloud has some good uses, in other cases where a solid infrastructure is already in place, it does not. So who wins the debate? For the moment, I am calling a jump ball. There are use cases for both and each organization should evaluate its own situation.