Managing Cloud Expenses Across a Hybrid Cloud Environment
Mastering the Tangled Web of Cloud Expense Management
In a recent blog, we discussed the charge and discount reservation variables associated with managing expenses for public cloud service providers. Here, we will explain the variables and inter-dependencies that should be taken into consideration when developing a comprehensive cloud expense management program that encompasses both private and public cloud services.
Unlike traditional telecommunication services, there are often many inter-dependencies of cloud services spanning across multiple cloud service providers (CSPs) that need to be properly documented in the configuration management database (CMDB) inventory, and then monitored and measured in order to make sure organizations have sufficient insight and control of all their cloud expenses. Before we get into the details, let’s first discuss the fundamentals of public and private cloud.
What is Public and Private Cloud Computing?
It is fairly common today for the modern digital worker to access many of their network, storage, computing, (commonly referred to as IaaS), telecommunications, (i.e. UCaaS) and business applications, (i.e. SaaS) services from remote servers over the Internet or via corporate WAN. A typical digital worker uses a combination of desktop, laptop and mobile devices to conduct their daily activities. The generalized term “cloud computing” is often used to define the ability to access these on-demand services that provide auto expanding capacity and do not require the need for direct active technical system management by the end user. As we stated previously, surveys conducted by Calero-MDSL clearly indicate that practically every organization implements a multicloud environment. The term multicloud is used to define when the organization relies on services from multiple CSPs that are then integrated into a single heterogeneous architecture by the IT department in order to provide ubiquitous on-demand services for their end users.
An example of a multicloud deployment is when an enterprise concurrently uses multiple cloud service providers to support their infrastructure (i.e., AWS, Azure, GCS, etc.), software (i.e., ServiceNow, Workday, Salesforce, etc.) and development environment (i.e., Oracle, AWS, Azure, GCP, etc.) business requirements.
In the world of digital services, cloud communication functions operate as a portfolio of applications integrated into and across one or more cloud service infrastructures. Organizations also often rely on different communication service providers to support various elements of their unified communication services (i.e., RingCentral, 8x8, Vonage, Verizon, WebEx, Windstream, Zoom, etc.) in conjunction with one or more communication platform service providers (i.e., Twilio, Bandwidth, Avaya) to support their unique business requirements and customize workflow messaging services.
The Difference Between Private and Public Cloud
Both private and public cloud frameworks are architected to be highly scalable and self-managed. Although both public and private clouds use a similar computing model to host services, they have several key differences when it comes to managing expenses.
Private cloud management is exclusively handled by the corporate IT department, which provides the service. In most cases, the enterprise owns all the hardware and network necessary to operate the private cloud. Cloud services are only shared within the enterprise and it is completely controlled by the company. The IT shared service extends their corporate private cloud by adding more servers to improve the performance and capacity. Similarly, to public cloud, the users are able to self-manage the capabilities and adjust the capacities of their respective cloud services.
The main difference between public and private is that a private cloud is always deployed within the corporate firewall and WAN, whereas a public cloud is typically accessed over the internet. In private clouds, the enterprise has greater control over the level of isolation, security and privacy of data. Unlike private cloud, public cloud does not require a large capital investment. Instead usage is metered based on different rate scales such as reserved, on-demand or, in some cases, spot. In both cloud scenarios, the TEM organization needs to implement ways to properly allocate charges to the consuming cost center.
What is a Hybrid Cloud?
A hybrid cloud environment is an infrastructure that interoperates the exchange of data and the lifecycle of workflows across the combination of a private cloud managed by the enterprise and one or more public cloud service providers. In a hybrid cloud environment, the enterprise chooses which business processes shall be run from the public cloud service providers and which business processes must remain running in the private cloud. Deployment decisions are contingent on many factors including data compliance, security, performance, cost and accessibility.
In addition to the inter-dependencies of cloud service providers, organizations often deploy edge computing as a means to bring computation and data storage closer to the location where it is needed, to improve response times and conserve network bandwidth. Each element of a hybrid cloud environment contributes to some aspect of a cloud service charge variable.
Private Data Center Expense Variables
The complete private cloud infrastructure is often hosted across several geographically distributed data centers. Similar to TEM, the IT expense management program must track ownership for all the corporate inventory and then process and allocate the charges from services across all the private and public data centers to the appropriate cost center. The types of charges that are often included in a private data center invoice are:
- Rent: recurring charges that represent the leasing commitments to host servers
- Power: encompasses the variable usage charges related to invoiced consumption
- Professional Services: represents the variable labor charges related to consulting, moves, adds, changes and disconnects performed by the data center personnel that houses the corporate systems
- Cross Connects: includes the recurring and non-recurring charges related to the data circuits and rent of physical cables and circuits that inter-connect servers within a data center and, in some cases, across different data centers
Analyzing Private Cloud Service Expenses
Similar to TEM, a fundamental requirement of cloud expense management for private data centers is to have an enterprise-wide inventory system that manages the assignment and financial ownership for each service. The sections below outline some of the essential information along with guidelines for monitoring, measuring and analyzing all the CMDB inventory elements that comprise an expense management program for a hybrid cloud environment.
- Data center inventory:
- A structured taxonomy for the types of servers
- Identification for each specific server deployed to each data center
- Configuration for each specific server
- Total quantities of servers per data center
- Parent & child relationships (e., cross-connects associated to each server and storage)
- Wide-Area-Network (WAN) inter-dependencies with terminating locations
- Services invoiced from each supplier (e., 3rd party leasing company, power company, etc.)
- Assigned contracted threshold agreements related to rate structures
- Enhance data center service invoice charges by:
- Structured vendor list
- Classification to the type of charges – recurring, non-recurring, usage and taxes
- Classification to the types of services – rent, power, professional services and cross connects
- Benchmark to data center characteristics:
- Analyze trends
- Identify anomalies
- Service business purpose
- Location (data center address)
- Cross connect service bandwidth
- Invoice date
- Transactional history with related financial information
- service orders – activation, moves, changes, disconnects
- activation, deactivation, suspended, reassignments
- Hierarchical perspectives of data center monthly invoiced charges
- By data center title
- By data center location to include longitude / latitude and Master Street Address Guide
- By global region, country, state, city, specific data center building
- Identification of network, (cross-connect) redundancy and points of failure
- Missing data center location information
- Tenant suite (when applicable)
- Service type – rent, power, professional service or cross connect
- Monthly breakdown of all invoiced charges by charge categories
- Taxonomy of recurring charges, non-recurring charges, usage charges and taxes
- Clarification of monthly percentage changes by category
- Identification of percentage changes in monthly charges by category
- Combination benchmarks to budget thresholds
- Total threshold spends to budget
- Total kilowatt consumption to forecasted kilowatts
- Identification of threshold overages, penalties
- Average data center cost by measured dimension (e., square feet, quantity of servers, etc.)
- High points, lowest points, averages, medium, median for each vector of charges
- Forecast out 12-month pro forma forecast based on 13-month rolling average
An effective cloud expense management program unifies the information from all data sources including carriers, cloud service providers, hosted data centers and other technology suppliers.
To stay on top of the latest industry news and topics, be sure to subscribe to our blog.