Back in the deep, dark past, when I designed and optimised relational databases for a living, I pretty much used every data modelling tool on the planet – at least that’s the way it seemed. They all varied in terms of strengths and weaknesses, but one that I developed a lot of affection for was ERwin, which back then was provided by a company called Logic Works. 

As the Wikipedia entry kindly reminded me:

“In 1998, Logic Works was acquired by Platinum Technology for $174.8 million in stock, which was in turn acquired by Computer Associates the next year.”

By then, I had moved on into the world of ERP, but a few days ago, a group of us from Freeform Dynamics had the pleasure of speaking with the guys at CA Technologies now responsible for the latest incarnation of ERwin – now known as CA ERwin.

Apart from nostalgia, I did have a good reason for wanting to catch up with my old friend, and that’s because of the recent announcement of CA ERwin support for SQL Azure, Microsoft’s cloud based RDBMS. 

We hear so much about rapid development in the cloud, but all too often the implication is that it’s legitimate for traditional rigour to be sacrificed on the altar of time to deployment. Indeed, some cloud environments, and those that advocate them, almost encourage a quick and dirty approach to development, with minimal analysis and design. The truth is that this will always come back to bite for any significant system, whether in the form of poor performance, lack of flexibility, or high operational overheads. 

Returning to CA ERwin, one of the traditional strengths of the tool throughout its evolution has been the ability to define a logical data model then map it onto multiple physical models. This is not a unique capability, but the ERwin line has always provided functionality that is both comprehensive and easy to use – ideal for professional analysts and designers who want to do things properly, but still want to move quickly in an unencumbered manner. It’s a philosophy that would seem to be particularly relevant to those wishing to take advantage of rapid cloud development and deployment, but without sacrificing structure and discipline.

To existing CA ERwin users, SQL Azure becomes just another supported database environment, but an Azure specific version has been announced by CA Technologies for those wishing to adopt a more structured approach to data modelling in a pure cloud environment. In related announcements, CA Technologies has also highlighted a portal facility to allow models to be presented and visualised by different types of user, and increased collaboration capability to encourage more reuse and sharing of models and components. Going hand in hand with the latter is a new concurrent licensing model

Standing back from these announcements, however, my overriding view is that anything encouraging and/or facilitating developers to apply structure and discipline to cloud based projects is to be welcomed. Adoption of cloud should not be an excuse for throwing analysis and design principles out of the window.

Dale Vile, Freeform Dynamics

Last week was extremely busy, so the first I heard of Amazon’s Direct Connect service being available in the UK was when my colleague Tony Lock forwarded me a journalist request to provide an analyst comment. In a nutshell, Direct Connect allows an Amazon customer to hook up to the AWS cloud service via a dedicated comms link, rather than having to send all traffic over the public internet.


The journalist asking for input was Rosalie Marshall at Incisive, and the article she produced can be seen here – worth a read if you want a quick summary.


The comment I provided, which is quoted in the article, was as follows:


“Relying on the public internet for core application connectivity introduces a degree of variability and uncertainty around bandwidth, speed and latency that is unacceptable to many large organisations, which are increasingly putting the emphasis on end-to-end quality of service management. Utilising dedicated links to cloud providers overcomes this and hooking up via incumbent communications service providers can also have benefits in terms of costs, monitoring, troubleshooting and support."


“While security, per se, should not be an issue when sending traffic over the public internet, provided it is appropriately encrypted, directly connecting to the cloud provider does take away a commonly perceived risk, which may make it easier to get sign off from non-technical stakeholders when making cloud-related decisions.”


These comments were based on various conversations with senior IT decision makers, along with, of course, insights from the extensive primary research we have carried out to explore the practicalities of cloud adoption. If you are interested in seeing some of this, a particularly relevant report is one that Freeform's Andy Buss put together a few months ago, entitled: “Cloud Connectivity; Carefully does it”, which can be downloaded from here.


You can check out that report at your leisure, but suffice it to say that one of Andy’s main conclusions from the research was that connecting to cloud services is a whole different ball game to enabling remote access. Just because you have the comms in place to handle the latter, doesn’t mean they will be up to dealing with the former.


Back to the AWS announcement, Andy later followed up with Amazon and arranged for the team here to speak with Robin Meehan, Chief Technology Officer at Smart421, Amazon’s launch partner for Direct Connect in the UK. Robin pretty much reiterated the points outlined above in my initial take, but we also covered some of the practicality.


Robin highlighted the importance of a one-stop shop for the entire service end-to-end (connectivity and AWS infrastructure services), pointing out that most enterprise customers want to use a specialist to outsource these kinds of activities as they are not core business.


This makes absolute sense. Picking up on the trend towards end to end service management in the enterprise space, one of the frequent snags is how to deal with parts of the chain for which you may not have the specialist skills in house – particularly for elements that are physically outside of the datacentre. More and more, there is a need for trusted partners to whom responsibility can be delegated, and that often means working with suppliers that offer a broader scope and more coherent service.


As Robin says:
 “We have deep connectivity skills and reach, as well as the application layer/IaaS skills, so when the customer says 'I can’t reach my Amazon EC2 instance', we are able to triage the problem effectively as we understand the entire architecture. For example, if it turns out to be an EC2 security group issue (aka firewall at the AWS end), we won’t blame the network.”

Of course none of this precludes Amazon customers piecing together the solution themselves, using their own expertise and general comms service providers, but as our research has highlighted, setting up the comms for business critical cloud services is not necessarily as easy as many make it out to be, particularly when more demanding applications and/or larger user bases are involved.


Anyway, the bottom line is that this recent announcement is welcome as it provides UK AWS users with choice that’s been available to US customers for a while now.

Dale Vile, CEO, Freeform Dynamics

I am currently looking at the extent to which it makes sense for those in the IT channel - VARs, resellers, ISVs, SIs and so on - to get involved with various cloud computing options.

I have an opinion piece lined up for Computing sister title CRN (should be published in the next couple of weeks) that outlines my current thoughts, which are basically that there is too much uncertainly at the moment to ‘bet your business' on cloud, but that it makes sense to start gaining some experience sooner rather than later if you can do that without distracting from your core business that much.

This is a bit of a generalisation, of course, but with all the rhetoric around channel partners having to ‘get with the programme' or risk being left out in the cold, I thought something was required to balance the debate.

The truth is that demand for most kinds of cloud based solutions is still pretty limited when you compare it to the IT market as a whole.

A lot of the numbers sound impressive when you hear them in isolation, but in relative terms, I am hearing from many of those that are active that it's still quite hard to find, qualify and close cloud related opportunities (compared to traditional business).

And from a business planning perspective, there is still the difficulty that the street price for many services has yet to settle. Some tell me that prices and margins can only come down, which makes ROI assessment quite hard to deal with.

Others say the margins on many of the products they sell are slim anyway, so price/margin volatility doesn't matter as they would look to do the same with cloud as they have always done - i.e. make money from spin-off revenue lines such as professional services and cross/up-sells.

To illustrate how inconsistent things are at the moment, I have also heard the view that some prices will actually have to go up as there is no margin in some of the early market offerings.

I guess if the service provider set their prices originally based on a direct Web-based sales and deployment model, then only later figured out that selling around the channel was very difficult (particularly in SMB), they may have ended up with numbers that don't add up in the indirect sales context.

Beyond demand and the ability to make profit, the one big concern I am hearing quite frequently is around customer ownership, but it's more of a question than a problem statement at the moment.

The issue is whether the channel partner would be expected to sign the customer on the service provider's paper, thus notionally passing control of the relationship to the provider.

Given that some cloud providers have little or no experience with the channel, and that lack of experience often translates to lack of discipline (e.g. in preventing/managing sales conflict in the field), the issue of whether providers can be trusted is clearly at the back of some people's minds.

Anyway, these are some of the things that are coming up in my conversations in addition to the usual discussion around the challenges of modifying your accounting models and remuneration plans to deal with the switch to an annuity based revenue model, but I am still very much in investigation mode.

So, if you have experiences, thoughts or opinions in this area, especially if you are working in the IT channel, I would be really interested in hearing your views. My contact details are on our website, or you could post a comment below, but the easiest way might be just to ping me an email here.

Dale Vile is managing director of Freeform Dynamics

Some back-to-front thinking in evidence?
 
One of the most frequent concerns about cloud is security. Andy Buss and I were discussing this the other day as part of a research scoping exercise. We are currently designing a study to look at the risk related aspects of Software as a Service (SaaS).

Something we always try to avoid in our research and analysis is falling into trap of generalising too much. In this case, it was important to acknowledge that businesses vary significantly in terms of their risk sensitivity, e.g. based on the degree to which they are regulated, the amount of confidential or personal information they handle, their operational dependency on IT, and their general risk awareness. Attitudes to security therefore range from extreme paranoia at one end to total complacency on the other. And even within a given organisation, some systems and data will be regarded as highly sensitive, and others will not.

The logic then goes that categorisation of applications based on their risk profile is a good place to start when considering which requirements lend themselves more to cloud based deployment from a security perspective.

So far so good, but then the conversation with Andy got really interesting.

The unspoken working assumption to this point was that application profiling would allow organisations to identify low risk candidates for initial cloud activity. To put it another way, if you're concerned about the hosted services model representing a security risk, then gain some initial experience with less sensitive applications for which security is less of a consideration.

The trouble is that for many small and mid-sized businesses, it could be argued that such advice would be flawed. Whatever the current perception, the reality is that a reputable service provider will almost certainly be able to manage application access and information security better than the majority of smaller businesses (and arguably many larger ones), so data and transactions would probably be a lot more secure in a third party hosting environment. The reasoning here is not rocket science, even though it may not be obvious to many. Service providers have the economy of scale to justify investment in top-notch technology and skills in a way that SMBs, with the best will in the world, could only dream of.

As responsible analysts, perhaps we should therefore be turning the logic on its head and advising at least some organisations to prioritise putting their most sensitive rather than least sensitive applications and data sets in the cloud first. While the original opposite view might be intuitively appropriate, that would be simply pandering to ignorance and ill-considered prejudices.

As this point, I can almost hear the abuse being directed at whatever medium you are reading this on – “Bloody naïve ivory tower analysts that have never done a real day’s work in the real world giving us bloody stupid advice about putting our most sensitive data into the hands of ‘fly-by-night’ cloud upstarts? They should get themselves a proper job and stop writing such crap”.

Then again, maybe this line of reasoning has got you thinking, which to be honest, is all I am trying to achieve.

In the real world, of course, it’s not legitimate to provide sweeping advice like the above. But neither does it make sense for those in IT to make sweeping generalisations about whether cloud services are or aren’t a good idea, on security or any other grounds. The point is that it needs some thinking about, and sometimes the most obvious conclusion can prove to be incorrect in many scenarios.

It’s for these kinds of reasons that one of my other colleagues, Tony Lock, and I put together a paper entitled “Applied Cloud Computing: A practical guide to identifying the potential in your environment”. In many respects, this was a reaction to all of the generalised opinion we hear on both sides of the house, as both the evangelists and the sceptics are guilty of the same crime in this respect.

The reality is that it’s all about context, and what’s appropriate or meaningful in one situation could be a total non-starter in others. Then there are those grey areas where it’s difficult to call it either way. With this in mind, after almost a quarter of a century in IT, I am still waiting for an example of a technology or approach that is universally right or wrong regardless of the circumstances.

Meanwhile, if you are interested in a more practical treatment of cloud computing, including some down to earth thoughts about security, integration, management and the general impact on the IT department, you can download the abovementioned paper from here.

By Dale Vile, MD at Freeform Dynamics.