Transparency: a core tenet of the cloud

Transparency has long been considered a core tenet of what defines a hosted service as “cloud computing”.

iTnews editor Brett Winterford.

The first place we usually seek transparency relates to price. A genuine pay-per-use model should be granular, easy to understand and clearly documented for customers.

Most enterprise cloud computes fall over at this first hurdle. Scour the web sites of HP, IBM, CSC and Fujitsu for their cloud pricing and you might be disappointed.

Even those that provide pricing to the public often go to great lengths to obfuscate comparisons with competitors by using different pricing metrics.

Time and again, attempts to compare services on a like to like basis are made incredibly difficult - a tactic the cloud providers must have learned from the mobile network operators.

Where else do we seek transparency? Consider outage reports. Regardless of your feelings as to whether cloud is “commodity”, customers should be entitled to a detailed post-incident report when services have been unavailable, if not for the simple fact that the service provider should be bending over backwards to restore customer faith in their capacity to keep your services online.

But compare the post mortems offered by the likes of Amazon Web Services and Melbourne IT to those offered by Microsoft Office365 and again you’ll note that the industry’s definition of transparency varies wildly.

In this week’s cloud cover bulletin, you may have already read an introduction to US based hosting company SoftLayer. This company, I am happy to report, take transparency one step further, and I felt it worthy of a column.

Many service providers use the commoditised nature of the services as an excuse to show little transparency with regard to the location of customer data and applications.

“It’s just in the cloud, man” is the expectation here. But no matter how virtualised and fluid a service provider’s systems may be, a customer does have some right to know. As we have learned through research by Truman Hoyle, there is a big bag fall of legal considerations most enterprise and government customers need to consider with regards to the location of data.

SoftLayer, to my pleasant surprise, take the opposite approach. The company is more than willing to detail its various investments in data centres, points of presence and fibre networks.

The company’s CEO Lance Crosby is happy to lay out where these points of presence are, what fibre routes the company has leased capacity on, and where a customer’s data is hosted. Right down to the most granular of detail.

“In Amazon land, they don’t really tell you where you are, you are just in x availability zone,” he explained. “In our data centres, we tell you what location you are in, what pod you are in, what rack you are in. Our core tenants are the empowerment of customers and transparency to customers".

How is that helpful in a cloud context?

For one, it helps administrators plan a strategy around redundancy across multiple data centres or availability zones.

SoftLayer offers customers an API (application program interface) with over 1600 available calls to features in its control panel, so that customers have visibility and control of their workloads. Crosby claims that his larger competitors offer no more than 60.

Armed with this information and a knowledge of where SoftLayer has dedicated fibre paths, customers can more effectively troubleshoot problems.

When there is an outage, customers want to know whether the problem lies with their application, with equipment in the data centre, with SoftLayer’s network or whether it is a one-off problem at the endpoint.

George Karidis, chief strategy officer at SoftLayer said there remains a line between offering transparency and providing a managed service.

 “We’re not trying to outsource a customer’s IT staff, we are just giving them access to the infrastructure they need to run workloads,” he said.

“We don’t want to run the workloads for them or determine their security.

“We just give them control and transparency, versus the IBMs etc that don’t want customers to see where things are breaking.

"They don’t want customers to see where they might not be meeting the SLA, because that’s where their margin is. It’s a fundamentally different philosophy.”