Cloud Services Could Be Made More Reliable with Premium Payments

Tom Keane
4 min readSep 4, 2022

Today’s businesses rely on the cloud for almost everything. It delivers a variety of services, including databases, software, analytics, intelligence, and more. Without it, many companies would have a hard time operating.

But the cloud has its share of issues. Slow recovery time, outages and other reliability problems cause a break down in productivity. Now enterprises are stating they could provide better service for a premium of up to 111% over the base price.

Read on to find out what the premium service would entail.

The Study

A study was conducted looking at three scenarios for improving the efficiency of a simple WordPress website. It was required to load web pages within three seconds of the request. A Python simulation was used to change the bandwidth, and virtual machine demands to analyze their effects on cost in various circumstances.

The experiment was based on Amazon Web Service rates, but researchers state that the results indicate a general outcome as other services have similar structures and pricing models.

The study looked at improving the WordPress app efficiency in three different applications as follows:

● Providing backup of the VM hosting in the same availability zone

● Providing backup of the VM in a different availability zone in the same region

● Providing backup in separate cloud provider regions

A cloud provider available zone is a virtual data center. Several of them in one geographic location make up a region. VMs are software-based computers that run inside of other computers.

VMs are prone to becoming unresponsive while availability zones can also go down causing an unresponsive system. In some instances, a regional outage can occur, but this is rare.

The baseline service involved in the study, which did not offer any protection, totaled $217.38 per month. This covered only the cost of the bandwidth. With this system, if the VM failed and there was no backup, the customer would have to replace it before they could use their system and access their information again.

The protected service would deliver 99.5% availability and would only recover service if the availability drops below 99.5%. It would also compensate users for outages that last more than a day and a half for 29% of the monthly fee.

The Applications

In this section, we will look at the three different applications reviewed in the study and how they would work with the premium service.

Same-Zone Active Backup

A service that provides a load balancer and backs up the VM with a separate active VM in the same availability zone would not result in any downtime if the VM failed. The user would still be able to take advantage of the 99.5% availability. Compensation for outages lasting more than a day and a half would increase to 44% of the monthly cost. It would also result in added monthly expenses to cover the extra VM and load balancer totaling around $311, 43% more than the baseline price.

Active Backup in Two Zones in the Same Region

A service that backs up the VM with a separate active VM in a different availability zone would result in the same costs as a service that works in the same availability zone. It costs nothing more to put the VM is a separate zone, but it would improve the availability to 99.99%. Like the other plan, it would offer zero recovery time and come with a $311 expense that raises costs by 44%. Compensation rates would remain the same as well.

Active Backup in Separate Regions

This method involves setting up the app in two different regions. The app actively working in both areas is said to be the most efficient of all options.

The model requires there to be four active virtual machines to host the app. They would provide a virtual load balancer in each region, directing traffic to the VMs in the different zones. The load balancers would balance the system and improve efficiency so an outage would not be noticed by the user, nor would it require any action on behalf of the user.

The traffic to the load balancers would be directed by the Domain Name System (DNS). The DNS could be configured to work with the best load balancer based on its proximity to the balancer, the delay on the path to the balancer, and weighting policies. It can also detect when a load balancer is an unavailable and direct traffic to the other balancer.

However, the active backup in separate regions comes with its share of issues. For one, it relies on a balancing mechanism that must access stored IP addresses to function. If the address becomes unavailable users will be unable to use the system until the cache is updated with the addresses from the IP system. This would result in an outage that will last an indefinite amount of time.

It also comes with a higher price point. Costs go up to 111% over the premium baseline making for a total of $457.80. If an outage occurs that lasts for more than a day and a half, customers are entitled to 62% compensation for the service.

There is also a general warning that more resilient services may come with few reliable guarantees regarding efficiency and compensation for outages.

Conclusion

New research shows that cloud providers may soon have the capabilities to provide increased efficiency across availability zones at a relatively affordable price. This could mean that individuals and companies no longer have to worry about extensive downtimes. Will you be opting for this type of service when it becomes available?

--

--

Tom Keane

21-year Microsoft Developer and Manager | Seattle, Washington