| The problem with 'Cloud Computing' part II - 'The bigger they are, the harder they fall'
| By Salar Golestanian @ Sunday, May 01, 2011 :: 1:09 AM :: 759 Views :: 4 Comments ::
|Last week we had the biggest news in IT with Amazon EC2 going down and strange thing was that the competitors did not make a meal out of the story - They had a good reason. They are all heavily investing in it and this was their biggest PR nightmare. Over night the perception has changed about Cloud Computing and to top it all, how about Sony’s PlayStation Network? Of course you have (I Know it is not strictly the same kind of cloud as Amazon). But it is big and it was so thoroughly hacked this week that Sony had to shut it down indefinitely. Did you also know that Sony’s PS3 firmware is effectively wide open, because they made a hilariously stupid security mistake?
This is the trouble I talked about last month in my blog , I talked about some of the inherent problems with cloud computing. However, in my wildest dreams, I did not think that my reservations will come true in such a short time. Barely a month went by and we have had two major networks going down. In the name of the Cloud, I cannot resist but to quote the proverb: When it rains, it pours.
“This is a wake-up call for cloud computing,” said Matthew Eastwood, an analyst for the research firm IDC, using the term for accessing services and information in big data centres remotely over the Internet from anywhere, as if the services were in a cloud. “It will force a conversation in the industry.”
Amazon has thousands of corporate customers, from Pfizer and Netflix to large number of start-ups, whose businesses often live on Amazon Web Services. Those reporting service troubles included Foursquare, a location-based social networking site; Quora, a question-and-answer service; Reddit, a news-sharing site; and BigDoor, which makes game tools for Web publishers.
Amazon has data centres around the world, but the current problems have come from its big centre in Northern Virginia, near Dulles airport. The problems included being unable to access data, service interruptions and sites being shut down.
Other organizations heavily investing in Cloud Computing came up with PR statements to lessen the impact of the Amazon interruption that could theoretically frighten their own Customers. For example according to Lew Moorman, chief strategy officer of Rack-Space, a specialist in data centre services, was the computing equivalent of an airplane crash. It is a major episode with widespread damage. But airline travel, he noted, is still safer than travelling in a car analogous to cloud computing being safer than data centres run by individual companies.
IMHO the problem for Cloud Computing Investors is going to be huge. It is not just to worry about the loss of data - but it is the fact that when a small section of the cloud goes down, it can bring down the whole lot for an unpredictable period. So if all your business is running on the cloud; well - just close and go home and come back when the rain stops.
Add the kind of issue that Amazon has had with a security attack like Sony has had then to be honest that is the end of it as far as the rock solid impression the market has had until now.
Complex infrastructure will naturally have complex crashes
Companies who run large clusters of virtualized servers with extensive storage farms and networking backbones are bound to have big crashes that are difficult to predict or solve.
It would seem that EC2 was so complex and the troubleshooting was extremely difficult to locate source of the problem that they needed to explain it with a document with a 5,694-Word Summary of Cloud Outage.
For a redundant Systems to work well it means that they need to be in two locations at the same time. So the scaling challenge to replicate data to multiple locations must be addressed. This seems to be missing for a number of high-profile online businesses who are still having problems today. Sadly, they probably deserve what has happened.
Strange that other Cloud Providers have not been attacking Amazon:
I have noted that none of the other Cloud Providers haven’t been attacking Amazon or been trying to convert Amazon customers to come to them. This is mainly because Amazon has the most stable cloud infrastructure of them all. However, now we know that cloud Infrastructure by its nature is flaky and Microsoft, IBM, and others know that their systems are not very stable today. It could easily be them tomorrow that fall like Amazon has.
Also, as Amazon customers take note of the damage, they are likely to boost their redundancy planning with more systems, more storage, and therefore spend more money. Criticising Amazon is clearly bad for their own business as the impacted Amazon customers just might revert back to traditional hosting solutions. The fact is that at the end of the day all this would mean that if you are going to host your application on the cloud, well it is going to be much more expensive.
Clearly 'Cloud Computing' guarantees are completely worthless now:
Amazon used to give service guarantees that the EC2 instances within the single datacenter location were fully isolated and provided undertakings that systemic failures could not affect within their data centre location. Clearly the zones were not isolated and their undertakings are worthless, now and for the foreseeable future.
Some Smaller Cloud Computing Providers are going down.
Also I am on the watch out for some public cloud computing providers that are failing? For example "Iron Mountain Digital". So it looks that many of the smaller cloud computing providers are not getting the traction they anticipated and are closing their doors, including some of the older firms. This is largely due to providing a far too tactical solution in a world where strategic solutions are sought. Moreover, newer providers have learned to use other clouds, such as IaaS and PaaS clouds, as their platform, whereas the older providers built their services from scratch and are paying for their own private data centre spaces. The new generation of cloud providers based on cheaper back ends is pushing them out of business. This is not a good pattern as you need money in such an expensive industry to develop goo infrastructure.
So what is the answer?
Some suggest use of multiple data centres and multiple locations as well as multiple cloud providers - WOW! - this no doubt will double one's costs that are already much higher than traditional hosting – Sounds like they are just asking for adding more complexities and more that can go wrong. IMO we are going to need much more transparency than we have had so far from the cloud providers like Amazon. Amazon has been very secretive about their infrastructure. In fact they don’t allow Auditors in. I understand why as competitors will learn of their systems and processes and setup shop copying Amazon. But the recent news provides enough reasons to want transparency and if Amazon, Azure or other providers are not going to give it, then we just need to shop around for someone that does.
So to use cloud we need to understand Cloud as a piece of infrastructure which means we need to be able to carry out standard compliance and audit of the cloud infrastructure to verify how it works. IMHO if you want to trust a cloud then look for one that is a Open Source Products. The whole thing with Open Source is open and can be inspected. One would assume that if you want to call it Cloud Computing then what you need to know how it can remain in the cloud and not become a Fog!
To get a better understanding of the Cloud Computing Security Consideration, Australian Government Department of Defence has done a great document here.
In Part 3 of this subject I want to discuss some alternatives to the current up down approaches available to us for cloud computing and discuss possible Open Source Scenarios.