Improvements to Public Cloud Services: The Road to Cloud 2.0

Some time ago, the science and technology news media were flooded with a lot of reports on "charging per minute of cloud services", which also made its myth of characteristic flexibility seem to have been broken. These reports even exerted no pressure on Amazon As for other same cloud service providers. But what is the next step of development? Perhaps we can say that after seven years of cloud computing development, we have fully seen what cloud computing 1.0 is. Although there are some advantages over traditional mainframes, these advantages alone are still not enough.

What will Cloud 2.0 look like? Here are some tips for sharing Cloud 2.0:

1, on the CPU core, memory, independent disk space choice

When your IaaS provider lets you choose from a series of monotonous lists just to make their work easier than to bring more flexibility to your organization's business, Could computing really be called cloud computing? Really, think about it. How can a service provider force you to choose what is called the best service for your application or database? For most current IaaS providers, it's like buying a car - when you want a leather seat When they were in a chair, they said they were able to provide a car with a skylight, which you actually did not want. Why choose this IaaS platform, allowing you to pay for the resources you do not need. This is the cloud computing 1.0!

If your business is currently using cloud computing from public cloud service providers, try the following experiment:

Choose a larger server and check CPU utilization. Then check the memory utilization. Finally, check how much disk space your business actually uses. And divide the amount of space paid by your business by the amount of space actually used. Summing these three percentages and dividing them by 3 gives an average of three dimensions for each of your organization's servers. This is the percentage of money that your business is wasting on this virtual machine. The cloud computing 2.0 will bring your business more flexibility, truly only pay for your business server resources used.

Look for an IaaS provider that allows you to choose your own number of CPU cores, memory, and independent block storage disk space. This allows you to size your enterprise system to your specific needs without the need to try to put a large amount of workload in a small space.

2, the so-called scale is actually two dimensions, and we have forgotten the vertical scale

You can actually add hot-swappable memory capabilities without disrupting the running server for a long time. The concept of vertical scaling is not new to adding resources to an existing server. So why do not we all seem to forget that this is an option when we all start to shift the workload to cloud computing? The answer is simple: the first generation of cloud like Amazon can not do that, which is why Providers of cloud computing 1.0 are always forcing their customers to scale out because it's in their power to make a profit, but not the best choice for the applications and employees they manage.

Why should you limit your business to vendors that only allow you to scale by adding more VMs that are not suitable for your business? Extensive workloads can be achieved by simply adding more CPU cores or memory Instead of adding more virtual machines to the existing system for better processing.

Second-generation IaaS providers will be aware of this and offer standard capabilities that include vertical scaling without having to restart as their core product.

3, Achieve better and more stable performance with dedicated resources

The following is a common scene in the first generation of cloud computing. First, start up five virtual machines. Then, benchmark on all five virtual machines. Keep a good virtual machine to continue running, the other four shelves.

Why do people do this? First-generation IaaS performance inconsistencies due to over-provisioning (too many virtual CPUs on physical servers) and mismatches of various commodity hardware.

We were told that we needed to solve this problem with code or that we needed a creative deployment tool. But can we really solve these problems? The second generation of cloud computing in its virtual resource allocation will be more creative.

Dedicated virtual machines, from CPU cores and memory to dedicated virtual machines, utilize better hardware to provide better virtualization than cloud computing 1.0 from resource pools. This means helping customers achieve better and more consistent performance.

4, easy to use

What do you do when designing an application architecture and its components? Most people use tools such as PowerPoint or Visio to graphically represent the various components and use wires to display Their network connection or data flow. So, why are all major IaaS providers still using check boxes in the project inventory list, allowing you to imagine the connection status between these projects? Do not force customers to visualize the passive imaging components, but should be able to use clear Image to visual display.

The core audience for cloud computing 1.0 is the developers, those developers who are well-disciplined and think of the world as a set of abstract concepts that can be linked together. With global IT spending reaching about $ 4 trillion, public cloud revenue hit about $ 4 billion. We can foresee that there are still 99.9% of the remaining large market demand, need to serve a wider audience. Cloud 2.0 will show users an easy-to-use graphical user interface. In fact, we've seen this happen as people flock to the Apple Store to experience the Mac.

Why it has now reached the era of cloud computing 2.0

VCR has been replaced by the DVR. Windows, not DOS, has also made it possible to have a single computer on every desk in every home. Instead of using "Lycos" or "AltaVista" to search for anyone's information, Google now searches. We have seen more than once the pattern of reform: When the first generation of the product is created, the market soon there will be a new and improved products come out.

Cloud computing 1.0 is already in its final stages. We will certainly see improvements in the capabilities of public cloud services in the coming years or even shorter time frames. Billing per minute is a good place to start, but more flexible on-site vertical scaling without restarting can help your business achieve better and more consistent performance, with improved, easy-to-use visualization tool features , Will be the promise that cloud computing 2.0 will bring to us.