I read an article by Sharon Wagner “Cloud Wars In 2014: Amazon Versus Google And Other Follies” which included the below:
Looking back at Amazon Web Services, Google Compute Engine and Microsoft Azure outages in 2013, we have seen glitches here and there, but overall, the quality in terms of availability and response time remains pretty high. This means that from the enterprise’s perspective, 99.99% Vs. 99.999% availability will not be the deciding factor.
So what will be the decision criteria in 2014? All three giants (and others like Hewlett-Packard and IBM) have high market credibility for performance and availability. Assuming all major cloud providers will catch up to provide comprehensive computing, database and storage capabilities, differentiating between vendors will be tough. At the end of the day, choosing the right cloud provider will probably be down to cost.
Just imagine being able to automatically run a full security scan within minutes every time someone makes a firewall change, to ensure that there haven’t been any unexpected holes created, without needing to worry about it.
APIs are wonderful things, without them the ability to let others take advantage of your software or platform are significantly reduced, and especially when it comes to providing cloud based services they are essential. They do have some downsides though, in certain use cases, specifically:
Just imagine you have a security service that monitors all your infrastructure and applications in a cloud platform. It constantly checks for accidentally opened ports, new servers being provisioned that are unsecure, and lots more.
At the end of last year, we announced our intention to solve the commercial side of OpenStack by extending our cloud orchestration platform to support it. Now, we have officially joined the OpenStack community as a corporate sponsor.
What are we doing with OpenStack?
Millions of dollars are spent, and in some cases are completely wasted, developing services on OpenStack. Service providers don’t typically have the same technical and developmental resources, or funds available, as vendors such as IBM, HP and Dell, and as such this DIY approach isn’t working for them. Even those working with hardware vendors to deliver services are struggling to do it in fewer than 18 months. Flexiant wants to give service providers choice and help them commercialize their cloud services quickly. We will do this by offering a rapidly deployable, functionally rich, service provider ready solution, underpinned by OpenStack, as well as our own platform.
With the latest release of Flexiant Cloud Orchestrator 4.0, we have significantly enhanced the translation and internationalization capabilities to make it even easier for customers catering to multiple countries or markets.
We have achieved this by enabling service providers and their customers to choose which language they want to present and/or use the Flexiant UI in. This can be set on a user by user basis.
At first glance this may not seem that much, but to handle translation of all dialog boxes, menu options, error messages, and even context sensitive help is quite an undertaking. To then add the ability for service providers to easily add, modify or update these to suit their customer base adds to this complexity.
So I have been (un?)lucky enough to have experienced both sides of this argument, in the service provider world for the last 16 years which is why I thought I’d offer my thoughts on dedicated vs. virtual hardware.
Last week I offered four of the seven reasons why I believe virtual hardware is better than dedicated. In today’s blog post, I’m going to discuss the remaining three reasons.
5. Performance Overhead
One of the classic points that tends to be brought out at this point is how much overhead is lost by the hypervisor in the first place, and that itself can make it inefficient. Although there is of course an overhead to having a hypervisor, it is significantly low enough that savings easily outweigh this elsewhere in overall utilization efficiency etc. I accept that there are always edge cases that won’t conform to this, but we are not trying to cover 100% of all scenarios, merely the general hosting/service provider market.