Cloud services are not all sunny. Those who push cloud computing have a strong interest to ensure its success. The alternative of owning the technology and network may seem archaic, but older methods of computing and communications do have their value and in many cases the cloud is not right. Moving to the cloud has many issues.My two previous blogs dealt with security, compliance and the possible service provider implementations of cloud services, "The CaaS Cloud; Security and Compliance" and "UCaaS, UC in the Cloud". It is not my intent to stop cloud computing.
First, what is old is not necessarily obsolete. Consider the mainframes. They have survived much longer than the PC/server pundits ever imagined. IBM still has a thriving business in mainframes. There are still supercomputer competitions. BTW, Cray just won the last competition.
For most businesses, if it works and the capital costs have been written off, then cloud computing has to compete with the old technology maintenance and upkeep, which is usually cheaper and more stable than the cloud alternative.
The requirements of enterprises, manufacturers, educational institutions, government agencies and all the others are different. One size does not fit all. If this were true, we would all use the same business software, with only one vendor and no third party developers. Competition would be on price and operational costs, not product differences.
Do you think that most organizations that have data centers will close them down? Not likely. A reason may be the data center is not paid for yet so it has to be operational until the costs have been capitalized. There may be leases that have years to run.
Regulatory, security and compliance issues may be hindering the cloud implementation. How much responsibility will the cloud provider accept? What happens when the requirements are not met? Is the enterprise left holding the bag? I think that the contracts for these issues of cloud services will become very complex with the provider trying to absolve themselves of any liabilities. This will certainly make the enterprise reluctant to move their functions into the cloud.
Remember time sharing, computing in a cloud that was popular 30 years ago? So the cloud concept is not new. It operated under a different business model and had a different name, but it was still a service that was off site for computing. There was still a data center, but owned and operated by a third party.
The more unique functions and features that an enterprise needs, the more likely they will not be found in the cloud. Where would be the profit for such services, unless the customer was very large and the investment were worthwhile? There will always be functions that are common to most enterprises. These will be the candidates for cloud services.
I believe that large enterprises will keep functions in-house and will only work with cloud computing when the scope of the service needs to be nearly ubiquitous and the users are mobile. The smaller enterprises that can not afford all the functionality of a large enterprise will be more attracted to cloud services because otherwise the smaller enterprise will not have access to these functions. There will probably be multiple service offerings by vertical market.
The cloud provider will set the Acceptable Use Policies (AUP). The AUP will favor the provider's business model and revenue. Read your AUP for ISP service to get an idea of the unbalanced arrangement that favors the providers.
The Service Level Agreement (SLA) will be measured over a long period of time, possibly weeks. Of course the SLA will be met when no one is using the service. The SLA is most important when the busy hour occurs. Experience with the SLAs of MPLS services is an example of the biased arrangements that clearly need to be renegotiated to satisfy the busy hour of the day.
The demarc for accessing the service may be much further away than the entrprise expects. The distant demarc means that the provider is not responsible to meet the SLA at or near the customer premises or desktop.
Check into the reports produced by the provider. Do they offer real insight into the usage? Are they demonstrating how to tailor the offering to meet the customer's needs after the service is turned on? As more features and functions are introduced, the customer should be able to evaluate who uses what and how well so that only the features and functions of value continue.
The cloud services can be offshore. What does this mean to the entrprise? This is very much the same as offshore outsourcing. The issues are the same. If the enterprise did not consider outsourcing, then what is the attraction of cloud services?
There will be several business models for the cloud providers. The business model may complicate some and maybe all of the issues mentioned. Here are the possible models I have discovered:
* A total cloud service where the provider owns the hardware, software, network and has the staff that implements the service.
* The cloud service can be a collection of dedicated or shared servers that run customer owned software. Here the SLA covers the platforms and network but not the features and functions offered.
* A software vendor operates on another provider's hardware and network.
* A reseller that owns nothing but sells cloud services for one or more providers.
The business model will have a great influence on the SLAs and AUPs that an enterprise will encounter. The stability of the service may be in jeopardy if the service provider business model is not successful. What if the cloud provider goes out of business? What if the cloud provider decides to terminate some function and features? The enterprise should have a backup plan in place.
