Interview with David Coallier of CloudSplit

Cloud Split
Cloud Split

Last month’s IT news seemed filled with articles about the possible failure of cloud computing and questions whether the cloud is the best place for all our data.

With the very public Sidekick customer data loss, hiccups with Gmail and the slow expansion of Microsoft Azure , what is happening?

I don’t feel that it is the end of cloud computing but as an avid user I felt there must be more to this issue. I was lucky enough to be able to interview David Coallier, the CTO of Echolibre, who was recently featured at TechCrunch50.

David is the lead architect on CloudSplit, their real-time cloud analytics solution for real-time analysis of your hosted solution. Because of their unique product, David has a unique perspective on what is happening in the cloud.

David Coallier
David Coallier

Q: How can products like CloudSplit assist with choice of cloud solution?
A: CloudSplit helps you with the analysis of your costs per application and per servers that are using “the cloud”. With a thorough analysis of your costs, you are able to determine which service you are most likely to save money with. Of course, costs are not your only issues when it comes to cloud computing; you need to have a reliable service with a solid reputation and a solid technical team that you can trust with your data.

Of course, CloudSplit can’t really take into account the reputation factor of each services, but in future versions we’ll give you the ability to compare the costs with how much it would cost you if you were using another service.

Q: Advantages to business of comparison shopping between cloud services?
A: This is a rather tricky question because most cloud services seem to be offering nearly the same prices.  However, when looking a bit closer, some are cheaper for people that require more outgoing bandwidth and some are cheaper when requiring more incoming bandwidth. One thing that seems to be often forgotten when people are comparing cloud services is also the actual availability of developers and documentation.

Cloud services will lower your costs however if you decide to use a service that has no library to interact with, your fixed costs will rise as you will need to develop and experiment with your own staff whereas if you use a cloud service with an already existing knowledge base, vibrant community and various libraries and types of documentation, your development team is going to greatly improve their productivity thus reducing your costs.

Q: Is the cloud ready for external/existing data standards?
A: I strongly feel about standards. Standards should definitely implemented in cloud computing, and from a personal perspective, the standards I would like to see is not for the people using the “cloud” but for the cloud providers themselves.

International standards on security policies, data protection, data recovery should be, in my very humble opinion, immediate standards to be addressed and enforced to all vendors of cloud services.

I encourage initiatives as such as the OCC (Open Cloud Consortium). However, until all major cloud computing players are members of such group, I don’t realistically believe the standards are going to be implemented.

Q: Who appears to be the fastest moving in the cloud environment?
A: I would say Amazon. They have been there since the beginning and seem to be the driving force behind new initiatives in the cloud world.

Q: Do you see new entrants to the cloud marketplace or an emergence of virtual resellers?
A: I do. In fact, I can see a few new providers coming in. However, I reckon the biggest thing that is to come is the utilities attached with cloud computing. CloudSplit is a perfect example of where the cloud computing marketplace is going to bring. We are approaching the market with an inexistent offer on a widely used service. Rightscale are an example of where the cloud computing market is going. We are going to see an emergence of ideas on how to use cloud services more effectively than the original offers by the vendors.

Q: Do you see more importance around aspects of data portability between services?
A: As a developer and data portability advocate, I definitely see the importance of data portability across services. The ability to simply switch to another service is important for both customers and vendors. It gives customers a great way to switch to a better product, but it also gives the cloud vendors a way to offer a superior service and let the customers easily switch to their product. This will obviously create competition and an increase in the quality of services offered to the customers.

Q: Is offline backup of cloud hosting still important?
A: Backups are important. To me, it matters very little whether it is online or offline, as long as some sort of backup solution is present. There are already many good backup solutions and anyone using any sort of computing (cloud or not) should be prepared and have backup solutions set up.

Q: Did Sidekick failure show a weakness in cloud hosting?
A: Many people believe Sidekick was a failure in cloud computing. However, I personally believe the problem was not the technology but the implementation of the technology. The simple fact that there was no real-time backup solutions to recover the clients data was a massive oversight by the T-Mobile team.

It does show that people may rely on cloud computing for a bit too much and the fact that it’s a buzzword may blind certain people about the reality of hosting. Cloud computing will not solve your personal problems; it will not fix your faulty developers and racing conditions. It will help you achieve something but won’t achieve it by itself. By not considering the potential problems, you are bound to have unexpected results and I believe this is what happened in the case of Sidekick.

The problem did not lie only in the cloud hosting provider but also in the architects that designed the system to not be redundant and not have external backups of any sort.

Q: How could have T-Mobile reacted better?
A: I don’t recall seeing anything very specific regarding the T-Mobile’s reaction apart from a few random blog posts and some vague quotes. However, following standard IT best practices would have definitely kept them out of the deep waters. There was neither apparent failover solutions nor redundancy setup which indicates that something went very wrong in the technical chain of decisions at T-Mobile. The best way to react is definitely to do their due diligence and get back on their feet. Nothing else can be done really.
I appreciate the time taken by David to answer these questions in the busy time leading up to the launch of CloudSplit.