Cloud Computing: Hype vs. Reality

The first order of business is to get the marketing side of cloud computing out of the way. The term cloud computing is a recent branding effort for a something that’s been around for decades. SalesForce.com has been providing an Internet-based, i.e. cloud computing, CRM solution since 2000. Citrix has been providing remote access to software applications for 20 years.

Cloud computing is still somewhat of an ambiguous term but as it matures as a hardware, software, infrastructure model it will deprecate terms such as Software as a Service (SaaS) and Infrastructure as as Service (IaaS). There are also several types of cloud types - public, hybrid and private.

So why has something that has been around for so long now become worthy of re-branding? In my opinion, ubiquity of the Internet, the maturity of web-based interfacing technologies and the rapidly expanding blur between desktop and web are the major drivers for many believing that we are on the precipice of a major paradigm shift.

In my view, the primary concerns with moving to the cloud model are security and connectivity.

Moving sensitive data and intellectual property to a cloud provider is a risk to be considered carefully. An internal infrastructure can be completely isolated from the outside world where malicious activity is voracious to say the least. So at the very least, some businesses will always need and internal infrastructure.

Access to cloud providers requires Internet access and that inherently could be problematic. What happens when a user cannot access to the Internet? What happens when there’s a power outage at either the consumer’s or provider’s end? Without Internet access, what functionality will users lose from their desktop software? Some of these concerns are already issues and will likely grow over time.

You could also make a case that cloud computing implemented as the status quo across all industries and organization size - in it’s current form – could be considered a national economic and/or security risk. As cloud computing proliferates, resources from disparate cloud environments will be interwoven resulting in nested dependencies. Why invent a wheel when you can subscribe to one and integrate it into your automated processes? A benign or malicious act that triggers a negative event across the Internet could have significant economic consequences and/or national security consequences depending on its impact.

The counter argument to the aforementioned is using the electric power grid as an analogy to the emerging cloud. The electric power grid has been a reliable source of energy nationally for decades; one that we don’t really worry too much about.

This argument does not hold water.

Knowing our national security depends on it, the federal government heavily regulates the utilities industry and keeps a watchful eye on it. Cloud computing is currently not viewed or protected by the federal government in this way. If a ubiquitous cloud is envisioned then at some point it must have strong federal protection.

Cloud computing has also been touted as the return to dumb terminals with all software being cloud and subscription-based. I can’t imagine that happening. Much of this article was written on my vacation to Disney World in between jaunts to and from Magic Kingdom, Epcot, et al. Because I’m cheap, er, frugal, I opted to not pay the $50/week charge for Internet access provided by the resort. Thus, my writing was done offline using software local to my computer. Binaries installed locally is going to be a way of life for the foreseeable future albeit a bit neutered when the Internet is unavailable.

If you are reading this because you are interested in cloud computing and are considering how to leverage it for your organization then you are ahead of the curve and need not rush. Take your time, develop a long-term strategy (including governance policies), and implement that strategy in controlled phases.

Social Networking, Privacy, and Identity Verification

The value of total obscurity on the Internet is overestimated and is more destructive than constructive. Our fears about privacy infringement are largely irrational; at least in the context of social networking.

The ability to be totally obscure allows those bent on malicious intent to cloak themselves in an anonymous or fake identity – allowing them to say anything they want, about anyone they want, and without any repercussions.

It’s still rare, but sites are starting to require identity verification and that trend, in my opinion, is certain to continue. It has to. Identity verification provides for accountability and enforces a set of socially acceptable rules of etiquette just as they exist in the real world.

Most of us aren’t afraid to have our identification verified for credit cards, licenses, loans, job verifications, and many other situations. The large majority of us willingly provide our credit card information to purchase goods over the Internet.

Is caller ID a bad thing? Don’t most of us frown upon those that restrict the rest of us from seeing their identity when they call by setting their caller ID to private?

I just joined a social networking site called BestThinking.com that requires subscribers to have their identification verified. And you know what? I was not only willing but found it refreshing!

They do it in a very clever way with little personal info - nothing that you would consider a security risk to divulge – and a short multiple choice quiz of information about you to make sure you are really you. The bottom line is that social networking sites where those involved are willing to be known by their real identity adds credibility to discussions and opinions.

Remember the sense of foreboding people had when purchasing goods over the Internet was first introduced? As time went on we realized that our fears were irrational. Yes, there is identity theft and sometimes on a colossal level. That is what comes with forging a new frontier. It takes time to get it right. In spite of that, we continue to purchase goods over the Internet because we consider there to be considerably more value than risk.

We are now forging forward further into the Internet as a new frontier where a new level of identity verification will become increasingly more required and hence a more credible place to share information.

Do I think we should encourage legislation to force identity verification as a prerequisite to participating on the Internet? That would be a hearty No.

I do, however, think identity verification will become ubiquitous within this coming decade for the sheer reason that deep down we want to mold the virtual world into one that resembles, as closely as possible, the physical one in which we live.

Why?

In addition to basic familiarity and the comfort that brings, the physical world is more soundly implemented; especially with regards to established rules of engagement. The Internet adopting attributes of the physical world, which have been tested, modified, and enhanced over many millennia, will help to realize the potential of the virtual world. Identity verification brings the Internet one step closer to maturity and in this corner it’s a welcomed step.

I'm interested in your thoughts.

The Future of Crowdsourcing

Crowdsourcing, for those who are new to the term, is the concept of entities outsourcing problems to a large and typically unrelated group of people. For example, Netflix has an ongoing competition - open to anyone - to develop an algorithm to predict customers ratings based on their past ratings that bests Netflix's proprietary algorithm by 10%. The prize is a million dollars.

Crowdsourcing is not a new concept. The Longitude prize was an open competition established by Great Britain in the 18th century to solve the maritime problem of discerning longitude at sea. John Harrison was awarded the prize for his invention of the Chronometer. He wasn't treated well either. The dude solved the problem and was delayed the prize money for a full 30 years.

John Harrison's poor treatment by Great Britain illustrates one potential problem with crowdsourcing however there are others - little to no contracts, lack of continuity with contributors, potential lack of interest thus little to no participation, low to no wages, and risk of malicious intent.

The global recession has resulted in rapid growth of crowdsourcing due to two major factors:

1. It is often cheaper for companies to crowdsource solutions as opposed to directly hiring or contracting with professionals.
2. There are lots of people out of work so the pool of willing participants is high.

When the economy turns around will the resources currently involved in crowdsourcing dry up? Will the competition for intellectual property and time to market pressures move corporations back to more traditional methods that are more easily managed?

My answer to both questions is no.

With corporations having the ability to tap into a world population for ideas and solutions there is bound to be better results than with a small set of specialists. Crowdsourcing offers the possibility of tapping into brilliance without having to interview for that special person who will develop that next killer product.

As far as the contributor is concerned, crowdsourcing offers recognition, flexibility, collaboration, pay, and other self-satisfying attributes. The city of Los Angeles provided a survey to it's population asking questions such as "What services should be cut to balance the budget?" with a list of city services from which the constituent may choose. This type of crowdsourcing relies on non-monetary rewards but still has a high rate of participation.

One of the more interesting things to consider is how crowdsourcing will effect various occupations such as those in the creative design industries. When creativity is outsourced to the world, there is a potential for deleterious effects on wages and the number of permanent positions in those fields.

In my opinion, crowdsourcing, like social media, is in its Wild West phase. There will be significant movement and change along the way and its current incarnation will be unrecognizable 5-10 years from now.

As crowdsourcing models mature and become easier to manage the majority of us will be involved in some sort of crowdsourcing as an inherent part of our lifestyle. Just as I continue to manage my LinkedIn contacts and update my status on Facebook, I will likely also be contributing to my favorite crowdsourcing activities.

Kano Model for Prioritizing User Stories/Requirements

Regardless of whether you subscribe to Waterfall or Agile as your preferred methodology, some form or prioritization of requirements (Waterfall) or user stories (Agile) will take place.

One of the more difficult tasks is helping the user community to determine the varying degrees of importance of each requirement. One method to help with this process is the Kano Model.

The Kano Model is named after Professor Noriako Kano who in 1987 developed a theory for product development to classify features into five categories based on answers to questions about the specific features. Following are the five classifications.

Attractive	Delighted to have but unexpected.
One-Dimensional	Features customers compare with your competition.
Must-Be	A must have feature.
Indifferent	Neutral about the feature.
Reverse	The customer do not want the feature and actually expect the reverse of it.
Questionable	Indicates that the customer is unclear about the nature of the feature

The features are classified by asking the customer two questions – one functional and the other dysfunctional – to which the customer selects one of 5 possible answers.

The questions are:

1. How would you feel if the feature was present in the product?
2. How would you feel if the feature was absent from the product?

The answers from which they may choose are:

1. I would like that.
2. I require that.
3. I don’t care about that.
4. I can live with that.
5. I dislike that.

The initial reaction of most people when posed with the functional/dysfunctional question think “won’t the questions offset each other?” As it turns out most often they don’t.

An excellent example that I’ve heard in the past is with a milk carton that has a thermometer on the outside so the customer can see what temperature the milk is. One may select answer 1 I would like that as the functional answer but select answer 4 I can live with that for the dysfunctional question.

The answers are then compared to a matrix below to arrive at the classification. The letters in the middle of the matrix represent each of the classifications via its first letter, e.g. A=Attractive. To take our earlier example of the thermometer on the milk carton, the functional was Like and the Dysfunctional was Live With thus the matrix indicates that the classification is A=Attractive.

	Dysfunctional
		Like	Expect	Neutral	Live With	Dislike
Functional	Like	Q	A	A	A	O
	Expect	R	I	I	I	M
	Neutral	R	I	I	I	M
	Live With	R	I	I	I	M
	Dislike	R	R	R	R	Q

With the Kano Model, one is able to ask the customer two short and concise questions and ultimately gather critical data regarding the importance of the feature to the product. As a result, a development team can easily discern what’s mandatory, what’s nice to have, and where the land-mines are.

There are additional techniques that can be used in combination with the Kano Model to further prioritize requirements like weighting features but alas that will have to weight for another day.