“And because we understand these things we can optimise our technology. For example we can design and build our own network infrastructure that offers class-leading price performance, excellent manageability and very good flexibility which is again very difficult for most organisations to do because they don’t have the economic scale to make it worthwhile.”
Because we understand these things we can optimise our technology. For example, we can design and build our own network infrastructure that offers class-leading price performance, excellent manageability and very good flexibility, which is again very difficult for most organisations to do because they don’t have the economic scale to make it worthwhile, Ian Massingham, Amazon
Hardware usage
The company also drives efficiencies by looking at the hardware it uses to offer its services. It negotiates with semiconductor providers for example.
“Intel is a longstanding AWS technology partner and it fabricates CPUs specially for us that have specific characteristics. They’re optimised for use with our particular type of workload and also optimised for use within the kind of environment that we provide. This means we can get better price performance out of our CPU technology because we can optimise for our specific use case and environment, squeezing more cycles out of the same piece of silicon,” said Massingham.
As a matter of policy, AWS publishes its pricing publicly and encourages transparency when it comes to costs.
“All of our pricing is public. It’s available on the website and customers can also access it via an API, downloading pricing into their own tools. And there’s also an ecosystem of AWS Partners that use that pricing API to populate additional tools that customers can use for cost assessment, predictive analysis, cost recommendations and a variety of other functions as well,” said Massingham.
“We’re probably more transparent then many traditional vendors where they have this idea of list pricing which few people actually pay and pricing is not that actually that transparent. It really ought to be because presumably competitive providers have the link systems, so it shouldn’t be that hard for them to make their pricing accessible and transparent in the way that AWS pricing is.”
Easily understood
Massingham said that it is a key part of Amazon Web Services offering for costs to be easy to understand.
“I’m really interested in what’s in the interest of customers and it’s important to have the maximum level of transparency and the easiest mechanism by which customers can access pricing data. Obviously, we are trying to build our services in a way which meets the feedback and needs of the customers that want to use them and that’s what we tend to focus on,” he said.
Controlling costs
The issue of controlling cloud costs is likely to become all the more important as technology and take up evolves and we enter what is being called a multi-cloud era of adoption. According to Sachin Sony, senior manager for cloud strategy with Equinix, this is a natural evolution.
You can turn on and turn off cloud connections almost instantly so if you want to use Microsoft for storage and you want to use Amazon for its compute capabilities you can do so and then you can turn down your compute capabilities and allocate more bandwidth for storage, so it gives you that extra level of flexibility, Sachin Sony, Equinix
“Enterprises are looking to consume cloud services not just from one provider such as Amazon Web Services, but probably also from Microsoft and Google as well – in other words from multiple cloud service providers,” he said.
“The current model that is most prevalent in the industry is to consume cloud services over the public Internet, which is of course fraught with its own challenges from security to scale to pricing. It influences all these factors.”
“By coming into a data centre like Equinix, you can privately consume these services over a cross-connect or cloud exchange port and you can do this as services over private interconnection throughout the public internet so it’s much more secure, scalable and much more cost effective to customers,” said Sony.
According to Equinix, while it can use multiple models to work out costs for customers there are two main models that are most prevalent. In the first, enterprises consume cloud services from cloud service providers over a cross-connect and only pay for the copper wire or optical fibre that goes between their deployment within the data centre and the cloud service provider’s service, which typically sits in the same data centre.
The second model involves a cloud exchange port that sits within Equinix’s own data centres and connects out to multiple providers.
Cloud exchange port
“Once you have access to the port, you can open virtual connections to consume multiple cloud services, say from Amazon and Google at the same time. The big advantage of the latter is that you can consume multiple cloud services more cost effectively because you don’t need to have a separate cross-connect to each one,” said Sony.
“You can turn on and turn off cloud connections almost instantly so if you want to use Microsoft for storage and you want to use Amazon for its compute capabilities you can do so and then you can turn down your compute capabilities and allocate more bandwidth for storage, so it gives you that extra level of flexibility.”
According to Sony, customers typically pay for cross-connects on a monthly basis, at a rate of typically “a couple of hundred quid” while ports can be purchased on a usage-based model or on a flat fee model.
“What we do is provide connectivity to cloud storage providers or cloud infrastructure services providers and that is extremely transparent. There’s a flat monthly fee and it’s visible right up there at the outset. We don’t charge any commission from the cloud service providers to get them connectivity,” said Sony.
Subscribers 0
Fans 0
Followers 0
Followers