Usage Metering and Usage-Based Billing for the Cloud

The 48th IT Press Tour had the opportunity to meet with Puneet Gupta, CEO and Founder of Amberflo. Puneet had the opportunity to work at Amazon as they built out their metering service that serves as the foundation for more than 200 services today. Today, AWS cloud metering is processing two billion events per second with one metering platform.

Puneet's vision is to transition the world to a more fair and transparent business model in the form of usage-based pricing. He highlighted the evolution from owned software like SAP from 1980 to 2000 to subscriptions like Dropbox and Salesforce from 2000 to 2010 to the usage-based software like AWS, Stripe, and Snowflake today. He believes the next wave of software will be powered by usage-based pricing.

I asked Puneet why companies that had gone "all-in" with public cloud infrastructure were now moving to a hybrid multi-cloud infrastructure due to the unknown and uncontrollable costs in public clouds. Puneet attributed this to the company's inability to know how they are using the public cloud. You can't manage what you don't measure.

This led to our discussion on the importance of metering.

Metering Is a Heavy Lift

It’s a complex data infrastructure problem. Companies need to be 100% accurate with a decoupled system of record. They must have real-time usage metering before they can provide a detailed bill. This includes knowing the amount of compute instance hours, cloud storage, API calls, payload, duration, Lambda function duration, and the number of transactions on any cloud, system, or application resource.

Metering is the backbone of all cloud operations. Everything needs to go through a control plane of API-enabled IT data infrastructure. It’s all software. Companies must have the infrastructure to track usage accurately.

Metering, as opposed to observability and logging, has to be accurate. Log observability can drop a record or two. Metering cannot; that's a design condition of the technology. Metering tells you what is being used by who, regardless of price.

Item potency is also critical. Process your record once and once only. Do not process it twice. This is how to ensure accuracy.

Metering Accurately and at Scale

Amberflo's solution provides metering as a platform and billing as the first application. With meeting de-coupled, the platform ingests all events worthy of tracking enterprise-wide. The platform is able to track billions of events, track them for real-time usage metering, and provide real-time usage-based billing. The platform has two components.

The billing cloud includes:

• A product catalog to define product feature line items
• Pricing and rating for tiers, volumes, and units — both block and fixed 
• The ability to build and deploy pricing plans with accurate usage data from metering
• Real-time invoicing and billing with audit trails 
• Combined to produce revenue analytics with real-time cost reports via dashboards and APIs

The metering cloud provides:

• Usage metering using a standard metering API
• Meter collection via a scalable and reliable meter ingestion service
• Processing and aggregating indices and by time slice
• System of record
• Combine to produce real-time usage analytics reports via dashboards and APIs

Amberflo has flattened the cost curve for metering and billing data infrastructure. Other providers' cost grows exponentially as you scale and limits use cases. Amberflo's internal costs are 1 to 10% of their competitors. This price-to-performance ratio enables use cases that others cannot, and the advantage compounds with scale.

Key Takeaways

• Metering is decoupled from billing.
• Real-time analytics and billing provide transparency.
• Accuracy of metering and billing leads to greater trust

While Amberflo is selling its solution to cloud service providers, it seems to me anyone using public clouds should be monitoring their use to gain transparency and better manage their costs.