Warning, this is a deliberately provocative post.
I take exception to some of the self perpetuating myths that the cloud bloggers are continuing to expound. Things like this from Krishnan on cloud ave.
“If internet connectivity is the issue, we should focus on making existing technologies to solve the problem , like Gears, better. We should strive to force our ISPs upgrade the network to meet the demand. Using software is not a solution”
What a lot of trot. Statements like this show how fundamentally divorced from reality these bloggers are. Here’s why;
1) Cloud services are not new. Get over yourselves. Dennis Byron describes this very well here .
“Cloud computing is not "the next big thing." It is "the first big thing" finally done right in the sense that all the stars are now aligned to deliver on the promises of the 1960s.”
I would also add that network providers like Juniper, Alcatel and Cisco have been doing this virtualisation ‘stuff’ for 20 years. The wake up call to the cloud computing industry is that they can help you a lot in sorting this out because they’ve already been on this journey. They also look at the complexity of offering IaaS and see it as quite trivial compared to what they already do.
2) All these cloud services are based on software, it’s the thing you connect to at the end of the network.Be it fixed or mobile, there is still software there!
3) You cloud / SaaS guys have no chance of “forcing’ ISP’s” to do anything because you are economically irrelevant. That’s right irrelevant. IDC believes the total market for cloud computing in 2008 is $12bn, if you are lucky you might be $42bn in 2013. That compared to the $1trillion of global Telco revenue is inconsequential, its barely a day’s spend for the US bail out fund.
4) Cloud computing, as Krishnan and his peers envisage it is a parasite on the networks of these Telco’s. Krishnan to his credit at least recognises that there needs to be a network, many don’t. But what they don’t realise is that their current business model exist solely at the benevolence of these Telco’s. Google is exactly the same. The technology has existed for years that means the Telco’s, if they so wished, could block Adwords content. Google for all its talk of net neutrality knows this very well, for that reason they are diversifying their business as well as adding network capability themselves. There is also strong rumour within the Telco community that Google has financial arrangements in place to ensure that such blocking doesn’t occur. (no source, just rumor). This is the case for a very real reason. The internet (and all other networks) cost loads of money to run. The average Telco spends 10-15% of the revenue on capital. That’s $100-150bn spent per annum to provide this network. Spends of this nature need to be recouped, & cloud computing companies better understand this because they are going to be asked to pay their way, guaranteed!
5) My final point is that the internet cloud brigade seem to fundamentally misunderstand customer requirements. For cloud to become truly mainstream, the providers need to be able offer SLA’s. Not uptime guarantees, but end to end SLA’s. While the cloud pundits have forgotten this, customers certainly haven’t. This simply isn’t possible with the internet, it’s a massive shared pool and anything goes. But cloud computing delivered over private connections enables this to occur. This is a fundamental shift in cloud computing that simply must occur. It will have major implications too. It will shift the business model to service providers as the aggregators. It will enable hybrid internal – external clouds to happen. It will also regionalise the provision of services to accommodate local security laws.
This diagram shows my view of how IT will be segmented over the medium term. This segmentation reflects my points above. There will be folks that are ok with an internet model for a few services, but this will become a small subset of the cloud computing user base. In my opinion most will be in private and hybrid cloud space due to the enhanced security, SLA’s and control it offers. There are also strong economic benefits. Some things will never shift out of the firewall.