Working in a political office means I am privy to the sorts of sales pitches that lobbyists, industry and community groups are constantly pushing on politicians. It can be weird, informative, amusing and at times plain scary, and I’m really valuing the critical thinking subject I took at University to help me better assess everything that comes my way 🙂
Anyway, seeing my passion and expertise is around technology, I do try to keep across what is happening as much as I can. Most of the big ICT companies are pushing the cloud computing pitch extremely hard, but I’ve found the moment you ask many of them questions about privacy, data portability, data export & archival, open standards, interoperability and issues of jurisdiction, just to name a few, they seem to baulk.
I think there are certainly a lot of opportunities in ‘the cloud’, but I think there is a lot of hype around this topic and I wanted to jot down a few thoughts that I think people should take into consideration when looking into cloud computing strategies. This is not a highly technical overview, but rather a bit of a mythbuster for those without a technical background to help in navigating the hype.
Sam Johnston pointed out to me earlier tonight a useful basic approach to ensuring you get an open cloud service which provides for the interoperability, portability and strategic control you want to maintain when moving to the cloud. If you have any good resources about cloud computing, please add it to the comments 🙂
I also strongly recommend you read the Open Cloud Manifesto which talks about this issue in greater depth, and touches upon other elements to consider when moving to the cloud.
Where is the cloud?
The term cloud computing came from the idea of services being delivered over the Internet, because the Internet has traditionally been represented on network diagrams as, you guessed it, a cloud. Some people use the term as the new SOA (and for all those who had to deal with the onslaught of SOA hype, you may enjoy http://soafacts.com/) and cloud can mean pretty much anything, which is why it is important to clarify what your vendor is trying to sell you. After all, services running in the cloud are still running on servers somewhere, so moving stuff to the cloud is moving stuff to someone else’s infrastructure and hoping they do a better, cheaper job.
I’m not saying you shouldn’t use cloud computing, but you should be very careful to understand exactly what you are getting, and you should be strategic. Charles Stross fans will already be familiar with the idea of the separation of data and processing power, and the cloud can provide enormous processing power without you having to necessarily hand over the reins to your data or your technology strategy. Cloud computing is not an all or nothing option.
Personally I believe you should always choose the best of breed tool for the job, committing to open standards and interoperability, and then you can mash tech together for your exact best needs rather than shifting to and away from cumbersome large solutions that try to be everything, and end up doing nothing particularly well, but I’ll leave that for another blog post 🙂
Saving the environment?
Whilst there is certainly an argument to consolidating old and largely unused hardware to reduce your carbon footprint and electricity bills, moving things into the cloud does not magically reduce your carbon footprint to zero. As mentioned, there are still servers out there, so the environmental benefits can be calculated by how much better the vendor is at efficiently using their infrastructure, than you. Again, it is just worth investigating the detail to understand the actual environmental impact, if this is important to you. Remember, refrigeration is a big contributor to carbon emissions, so it isn’t just about the hardware 🙂
How much money can I save?
There are certainly some great opportunities to save money by using cloud computing for some of your systems. Often you can get online services that can be cheaper than the cost of maintaining and running your own systems. It might be worthwhile to consider the cost against that of shared services under your control though rather than looking straight to the “cloud”. For instance, in Australia there is a large amount of projects around government data centre consolidation, where some costs savings can be found but the data, software, infrastructure and strategy stays under their control.
It is also worth considering the exit cost of any new solutions. Can you get access to export your data at any time, is it safely archived somewhere you can access in the unlikely but possible case of your cloud provider folding, or a contract disagreement? Can you migrate your data/service from the cloud vendor to another vendor/solution relatively easily? These are all important considerations when faced with “the cloud will save you money”.
What about my data?
What format is your data stored in within the cloud? Physically where is the data and what are you legal obligations in relation to data? This is an important concern for government where you shouldn’t store particular data sets outside of your legal jurisdiction, and government departments and agencies often have quite stringent privacy and other obligations.
Can you get immediate access to the most recent data if the “cloud” dissipates (had to make a joke like this sometime, sorry)? Where is the data archived? If you can export your data, is it available in a format that other applications can use?
All these are important considerations, because if your data is being updated in the cloud, but is not truly retrievable, you have a real problem.
The silver lining
There are a lot of opportunities to be found in cloud computing and you will find many, many blogs and presentations espousing the benefits of cloud computing. I wanted to write a short blog post to help people consider some of the issues. If you choose to move some stuff into the cloud, you are choosing to hand over the keys to your most treasured possession, so you need to make sure you aren’t locking yourself out.
You aren’t powerless in this transaction. You need to know what you want, know your exit strategy, be sure that your cloud solution is open enough to be flexible and interoperable, be comfortable with how much control you are giving up, and be sure you retain enough control to meet your obligations.
If you are comfortable with all of this, you can engage confidently with cloud vendors and demand what you need rather than being content with what you are offered 🙂
niiiice, love the way that you think.
Remember the browser wars? The pc/mac/nix wars? Welcome to the cloud wars, this is just history repeating its self; the sh*’s about to hit the fan, consumers are going to be penned in, developers are going to be battling the overhead of standards spending 90% of their creative energies climbing walled gardens, and thus, the economy is going to suffer…
but
I see a bright future for open standards Pia,
The dogma vs grok wave has been collapsing for centuries and I just have a wonderful feeling that just like “the enlightenment”, the current nefarious tech high priests are losing the battle to lock us in, and are in fact not be able to stop the fire of realisation will spreading beyond the original foss Galileo’s once dubbed crazy. Everyones eyes are slowly opening to what’s in the magic black boxes we call software, and through this are seeing good from evil of their choices and understanding the repercussions short sighted tech choices bring. It’s a bright future.
I like the idea of the manifesto but upon reading I was appalled to see the document was written in “Spring 2009”, which I’d argue is worse than a spelling mistake in a resume.
this is no difference than outsourcing, one has to think clearly on the benefits and what one tries to achieve.
if u worry about your data and want full control, there is a private cloud for that. hosted data risks compromise & unintended exposure, mitigate it. most cloud providers provide tools to allow u access your data, if they don’t, move to providers that do.
that reminds me, we’re still trying to dictate acceptable list of countries for g00gle to store mails
cloud is just about getting standardize (sort of), it’s unrealistic to ask for standard compliance when u may not even care/know what it means. u can wait till all standards are sorted, however u may miss the bus by then.
it’s time to stop building infrastructure from ground up, start sharing what’s out there (be it infrastructure, software, or platform). http://csrc.nist.gov/groups/SNS/cloud-computing/. most cases u lose some and u win some.
as u said, one needs to think carefully whether/not to jump to the cloud (some say it’s very comfy sleeping in the cloud). is there a need for it? hope govt aint jumping into it cause everyone else says so or the vendors show great powerpoint on it, there ought to be complete audit and consolidation on where the govt has been wasting so much of tax $$$ (this implies funding projects too)
at the end of the day am looking forward for the ppl to embrace the cloud and help shape the cloud that we want…
I’m personally quite happy that my employer (Rackspace) sees their product as execution. Having proprietary systems and locking you in is not seen as productive. It’s a pretty exciting place in the industry to be in 🙂
I did a podcast interview on cloud computing a year or two ago for… ? Richard Chirgwin? at this site: http://www.itradio.com.au/networking
Perhaps you might try to track it down?
Here it is:
http://itradio.com.au/networking/?p=69
Thanks Brendan, tha’s really useful 🙂
You allude to, but do not expand upon a huge issue. The open cloud manifesto fails to mention it at all.
The issue is this: physically WHERE you data is, while within the cloud, can be a huge problem.
To explain why, I want to venture into a parallel issue in a related
discipline.
In photography, current Australian Copyright law says that the person who pressed the button on the camera is the copyright owner (modulo the usual work for hire stuff you expect with copyright). {{Professional photographers taking family portraits is a weird exception, fortunately not relevant here.}}
This means, no matter where the image resides, be it as a negative, a print, or a scan in a computer, the person who pushed the button is the copyright owner, because it was their “creative act”. The law was written this way to fix an earlier problem.
Prior to 1968, the Copyright Act was different. At that time, whoever owned the negative owned the copyright. This caused problems when dealing with photographers contracted for any kind of work, as opposed to employees who were never a problem. Newspapers and magazines had problems that the 1968 Act improved.
The photography rules apply to those photographs when the are used in Australia, even if they were taken outside the country. Conversely, Australian copyright law is irrelevant in any other jurisdiction, even if the photograph was taken here.
What does this have to do with cloud computing?
Simple: if your data resides on a server in an “own the negative, own the copyright” jurisdiction, they can legally appropriate your data, because in that jurisdiction THEY own the data, because they own the disks. The only time your data belongs to you is when it resides on a server in a “push the button, own the copyright” jurisdiction.
This makes me *very* wary about having my data magically transport itself around inside the cloud to where it is most “efficient”. How do I tell the cloud “only move my data to other bits of the cloud where I still own it”? Especially when the hype merchants repeatedly gloss over this elephant in the room.
(How do you tell the cloud to route your data around an “own the router, own the traffic” jurisdiction?)
Oh, and by the way, do you keep track of all the random people you ever asked to take a picture of you and your friends with your camera? They own the copyright. Did you get them to sign a copyright transfer to you? There are times when the law is an ass, but in I’d prefer my data stay here in Australia.
Thanks Peter, I had been thinking along those lines, but great that you could articulate the problem so nicely 🙂
Good work on this article. It should help folks to ask the right questions when they’re taking a closer look at cloud options.
I think the definition of cloud as services over the Internet, or outsourcing, is too narrow. I know, it’s vague already! But it’s hard to argue that private cloud systems like Eucalyptus should be excluded.
I tried to tackle a comprehensive definition in http://mdzlog.alcor.net/2010/04/22/on-cloud/
Hi Pia,
Thanks for a such an interesting forum last night at ANU. You really made me realise how little I know about something I use everyday, as well as the potential of the internet.
I was wondering if you could give me some info or send me in the right direction as far as setting up a website is concerned? I really have no idea where to start, but I thought that you would be the perfect person to ask. If you think you could help me get started, just shoot me an email and I’ll outline the kind of site I’m hoping to create etc.
Thanks again, and I hope we can get you back to talk about other interesting tech stuff in the future.
Thanks
Tas
Cloud computing is maturing fast enough to consider by ICT professionals or companies who may want to use the new technology. Amazon web services, Google Apps and One Ubuntu are only just a few of the companies who are now aggressively marketing their cloud computing stuff.
Plunging into this new technology is really quite scary especially if you are thinking about data privacy and disaster recovery.
I have been using the Google Apps Cloud computing since Google released it more than 2 years ago and it is really very helpful and now becoming the leading software for document collaboration management system in the cloud. But at the end of the day there is a very thin fine line for data privacy.