When Free Is Not Really Free: What Does It Cost to Run a Database Workload in the Cloud?

Abstract

Abstract. The current computing trend towards cloud-based Databaseas-a-Service (DaaS) as an alternative to traditional on-site relational database management systems (RDBMSs) has largely been driven by the perceived simplicity and cost-effectiveness of migrating to a DaaS. However, customers that are attracted to these DaaS alternatives may find that the range of different services and pricing options available to them add an unexpected level of complexity to their decision making. Cloud service pricing models are typically 'pay-as-you-go' in which the customer is charged based on resource usage such as CPU and memory utilization. Thus, customers considering different DaaS options must take into account how the performance and efficiency of the DaaS will ultimately impact their monthly bill. In this paper, we show that the current DaaS model can produce unpleasant surprises -for example, the case study that we present in this paper illustrates a scenario in which a DaaS service powered by a DBMS that has a lower hourly rate actually costs more to the end user than a DaaS service that is powered by another DBMS that charges a higher hourly rate. Thus, what we need is a method for the end-user to get an accurate estimate of the true costs that will be incurred without worrying about the nuances of how the DaaS operates. One potential solution to this problem is for DaaS providers to offer a new service called Benchmark as a Service (BaaS) where in the user provides the parameters of their workload and SLA requirements, and get a price quote.

Cloud-Native Databases: An Application Perspective

“…While the authors focus on AWS EC2, the DBaaS services of the same provider are not considered. A similar aim is conveyed in the work by Floratou et al albeit with a critical look at unpleasant surprises in terms of financial risks when using DBaaS [5]. The findings are that more expensive hourly services may turn out more cost-effective overall, which is substantiated with observations of MySQL and SQL Server running on local hardware.…”

“…The lack of a reusable testbed from the related work is evident. Szczyrbowski et al [13] Seriatos et al [12] Mian et al [9] Floratou et al [5] Bagui et al [1] Costa et al [3] We conclude that cloud-native databases are a challenging topic in need of more formal expressions concerning their configuration and characteristics and of more experiments. We suggest that future research should be directed towards a holistic approach of assessing flexible database options in the cloud which involve self-hosted data containers, blob storage services and DBaaS.…”

Enforcement and Spectrum Sharing: A Case Study of the 1695-1710 MHz Band

“…In [13], the authors examine cloud-based database services. The advantage of this costing approach is that operating costs are explicitly included.…”

A Consumer-Centric Market for Database Computation in the Cloud

“…Contrasting our work with Benchmarking as a Service (BaaS) [18] is meaningful when workload repetition is significant. We assume a consumer who repeate RDS and EMR workloads without any modifications, and with each repetition tested a different configuration; once all configurations were tested, the consumer would continue using the best one in subsequent repetitions.…”

“…Floratou et al [18] propose a Benchmark as a Service (BaaS) approach to help consumers select configurations. This new BaaS benchmarks user's workload and use the optimal configuration to execute the workload repetitively.…”