Software application development
This blog is intended to provoke discussion and exchange between like minded software application developers, engineers, architects, project managers - and keen hobbyists too.
Wednesday 12 November 2008, 8:47 AM
I wandered lonely as an unmanaged cloud service
Well this is the line that some web application performance monitoring vendors are seeking to take in an attempt to proffer forth their wares. Manage with our product - they say – and a bright new world of scalable cloud-facing IT deployment becomes possible securely managed right alongside your internal infrastructure.
IDC predicts that spending on cloud services will reach US$42 billion in the next five years and that it will capture 25 per cent of IT spending growth by 2012. To meet these growth expectations – says IDC - adopters of cloud services will need additional tools and updated deployment strategies to succeed.
According to the analyst firm’s latest ‘impartial’ (but possibly externally funded) survey respondents’ top concerns for adoption include performance level assurances that, “Support the need for new monitoring and management tools to help ensure the success of these deployments.”
Of course, cloud computing and virtualisation technologies are comparatively new beasts on the evolutionary computing scale – and data centre managers from Wolverhampton to Wisconsin are probably having sleepless nights over server sprawl and the precipitation slowly dripping down their necks from the company’s latest cloud computing directives and goals.
Moving from the fixed capacity (at any given time) of the data centre to the interstate multi-lane application traffic freeway on the cloud clearly demands management if it’s going to be cost effective and work well in terms of data throughput, capacity handling and still retain the ability to be fine tuned. But the solutions that vendors are producing to address these issues are being sold with what smacks slightly of ‘hey guys – this is all really new complex stuff so you need us’.
Whether I am being fair or not, companies in this space include Hyperic who say that their product can take server deployment time down from days to minutes. The company says that its product is currently managing over 3,500 VMware and XenServer virtualisation deployments. Also referred to as ‘private clouds’, these environments consist of both physical and virtual servers and typically support high rates of change as virtual servers are easily added, subtracted or moved to improve server utilisation and maintain service levels.
Once again with see this kind of thing, this is probably excellent technology, but I feel it’s guilty of a little oversell. I know the market is burgeoning and everyone wants to get a slice of the pie, but please… a little more gentle alto stratus and a little less hurricane force mushroom cloud please.
PreviousComments on this post
This comment has been deleted at the users request
After I posted this blog last week I received a number of mails from the aforementioned company Hyperic. I guess they could have put this information up themselves, but – as they didn’t – and I think it does add to the story, here’s what CEO Javier Soltero wanted to say in response to this blog:
“Certainly the technology industry has many memorable instances where companies tout new products on the back of a trend that do nothing more than create more work and more problems. Yet we cannot deny that cloud computing is changing the technology industry dramatically and updating the way in which companies use and pay for computing services. Vendors like us will see this as an opportunity to respond with new products that address the technical problems *and* are delivered in a manner that matches the opportunity.
“Thus, merely releasing a version of our product to work seamlessly in the cloud without the corresponding licensing, pricing and delivery changes would be half-baked.
We started by focusing on the problems unique to managing large scale web apps (different from traditional ‘systems management’), continued with the advent of virtualisation and is now moving to the next level of infrastructure technology, with cloud computing. It’s not about just focusing on the technology that is the ‘new kid on the block’ but taking a visionary view toward those that have the potential to change the state of the industry. In my opinion, exciting management software companies are the ones that are helping enable the adoption of new, disruptive trends in computing. Everyone else tends to be focused on trying to compete on either price or me-too functionality.”
Don't ya just love the marketing guys in the computing world? Cloud computing sounds really interesting, yet even despite having looked into the Azure toolkit, can't help feeling the world closing in around me in terms of lock in. Maybe this lock in isn't actually going to affect me, but if I create cloud applications I am certainly assisting the lock-in of my clients.
If I develop an application using Windows Azure, not only am I tying my clients to Windows, personally I have no issue with that, but I amd tying in my client directly to Windows and ultimately Microsoft. I do have concerns with where my clients data is stored, and usually they do too. So, the client asks where is the data stored then, and you reply 'in the cloud', how useful. Which continent is this cloud over then?
Customers generally don't like Grey Areas. Grey areas are what lawyers make megabucks from. Interestingly enough, it's winter here, all the clouds passing my window are grey areas.
Another thought, don't you just know that some company is going to come up with a cloud computing product called 'silver lining'?
You have a damn good point here.
Even if some companies would take "don't worry, your data is in the cloud" as a positive right now as it's still sexy and cool... let's remember that we live in a cynical old world and pretty soon someone will use the following headline:
Gloomy & Grey Outlook Predicted for Cloud Computing
AdrianB
All those wonderful vendors selling "cloud computing" services are just selling server access and what used to be called time-shared computer cycles. Lock-in is the mechanism that makes cloud computing pure BS. All you're doing is "outsourcing" your server functions to the vendor selling you the "cloud access".
I can already see that certain customers will get better quality access when they pay more money for the "cloud computing" services they buy. Whether the services are for internal to the company services or external customers the purchaser wants to attract, the level of access can easily be controlled and the billing can be "metered" at that point.
Look for ISPs to get into this in a big way when they figure out how to run the "taxi meter". Maybe they can modify that software they were going to use to control customer access to the Internet with?
