Evaluate Weigh the pros and cons of technologies, products and projects you are considering.

Azure B-series VMs reduce resource waste for bursty apps

Paying only for the resources they consume is the ultimate goal for many enterprises in the cloud. Azure's B-series VMs make that goal more attainable -- for certain workloads.

Targeted at workloads with variable compute demands, Azure B-series VMs offer more flexible -- or, according to...

Microsoft, "burstable" -- CPU performance. And while this VM type is cost-effective for certain workloads, it won't be a fit for all enterprise apps.

The cost challenge with cloud VMs

Most of the time, VM instances in the cloud do not run at full utilization. Many cloud workloads are variable, meaning they experience busy periods, in terms of demand, followed by a lull. For example, many websites see a spike in traffic first thing in the morning, at lunchtime and then again later on in the evening, while, the rest of the time, the server on which that website runs only consumes a fraction of the resources available to it. But cloud providers bill for those resources, regardless if they run at 10% or 80% utilization.

No one wants a slow website or database server. Users want top performance at all times, but on the other hand, paying for compute resources to just sit there is wasted money. Simply put, we still lack dynamic sizing flexibility around cloud VMs.

With its Azure B-series VMs, Microsoft provides a way to address the issue of wasted resources and nudges users closer to the holy grail of cloud computing: paying for the exact resources you consume.

An overview of B-series VMs

With the Azure B-series, users select a baseline CPU performance level for their VM and then accrue credits whenever that VM runs under that baseline level. This lets users bank CPU credits during low-consumption periods to use during high-consumption periods, when they need to scale up their CPU performance.

CPU credit accumulates at various rates, depending on the size and use of the VM. Users can choose to either accrue or consume their credits and view the status of their credits in the Azure portal.

As the VM runs, it can be in one of three states, and it accumulates or consumes credits accordingly:

  • A VM that runs under baseline performance accumulates credits.
  • A VM that runs above baseline performance consumes credits.
  • A VM that runs at baseline performance results in no change.

Credits accumulate on a per-minute basis, and each Azure B-series VM starts out with credits based on the formula: number of VM cores * 30.

Therefore, a single VM core has a credit value of 30, while a four-core VM has a credit value of 120. As the core count increases, the baseline CPU level increases as well. A single-core VM has a baseline CPU performance level of 10%, while a four-core VM has a baseline performance level of 90%.

Other factors to note

Apart from their usage model, actual B-series VMs are no different than any other Azure VM family in terms of their software and available hardware configurations. B-series VMs are not yet available in every Azure region.

These bursty Azure VM types seem to be in direct response to a similar type of VM from Amazon Web Services, called T2 instances. In both cases, users will see these VM types as a good fit for low-traffic websites and proofs of concept, especially when compared to the more traditional, always-on VM types.

This was last published in February 2018

Dig Deeper on Cloud computing pricing and economics

Join the conversation

1 comment

Send me notifications when other members comment.

By submitting you agree to receive email from TechTarget and its partners. If you reside outside of the United States, you consent to having your personal data transferred to and processed in the United States. Privacy

Please create a username to comment.

What workloads do you run on Azure B-series VMs?
Cancel

-ADS BY GOOGLE

SearchServerVirtualization

SearchVMware

SearchVirtualDesktop

SearchAWS

SearchDataCenter

SearchWindowsServer

SearchCRM

Close