Spending Your Money Judiciously on Cloud

Executive Summary

With the accelerated rate at which businesses are adopting Public Cloud today, it has become extremely important for businesses to adopt methodologies to manage their Cloud finances effectively. Companies today have a range of options to cut their costs in Cloud. For instance, they could take advantage of the state-of-the-art Cloud architecture patterns, use effective reporting tools, introduce automation solutions for inventory control and improve their Cloud Governance strategies.

This paper discusses the various approaches we provide to our customers to stay cost-effective in Public Cloud without compromising on the real benefits reaped from Public Cloud adoption.

Introduction

Cloud is now an integral part of every organization’s digital transformation roadmap. The term “zero CAPEX” has been frequently associated with Public Cloud platforms. Public Cloud vendors are globally expanding their footprint and exponentially increasing platform capacity each day. The true benefit of “economies of scale” is passed on to customers – As a result of this, the costs of Cloud services have been continuously decreasing and the number of services are on the rise.

Cloud Computing offers cost advantages primarily because as a user of Public Cloud, you are using space in a third-party vendor’s data center. You thereby save on real-estate, equipment, power, backup and manpower, which are offered to you as a service by your Cloud Provider. The immense benefit of the pay-as-you-go pricing scheme also makes Cloud financially attractive.

However, as a regular user of Cloud, a company can experience a phenomenon we call “Cost Explosion in Cloud”. This scenario can be compared to that of a shopping experience in a supermarket. When you purchase services in Cloud, you may end up buying items that you really do not need. As time goes by, these items that you over-purchased end-up fattening your Cloud bill. “What does my bill look like? How do I itemize my bill? How can I bring down my bill the next time I shop?” These are the queries that intrigue most businesses.

“26 percent of respondents to the Right Scale 2016 State of the Cloud Survey identified cloud cost management as a significant challenge, a steady increase each year from 18 percent in 2013.

Cloud cost management provides a significant opportunity for savings, since few companies are taking critical actions to optimize cloud costs, such as shutting down unused workloads or selecting lower-cost cloud or regions.”

In this paper, we discuss various measures an organization could adopt to control the spending on Cloud. This paper is organized into four key sections. In Section 2, we talk about the “What” aspect – we make you aware of what is it that you are spending on. Section 3 talks about the “Where” aspect – we help you see where exactly you are spending your money. In Section 4, we cover the “How” aspect – here, we help you with various solutions and techniques to lower your Cloud expenditure. Section 5 covers a few Case Studies with illustrations on how we helped our customers solve their cost problems. This paper has been written to be agnostic of any specific Cloud platform. However, we have used examples from the environments of various leading Cloud Providers, and the case studies use these platforms as the basis. Scope of discussion in this paper is restricted to Public Cloud.

 

Getting Timely Insight into Cloud Expenditure

According to Gartner research1, the Public Cloud market is estimated to reach USD 383 billion in year 2020. Every organization is investing significant time and money in understanding, experimenting and utilizing Cloud services. The reliance on Public Cloud service providers is only going to increase in future.

Any organization with a Public Cloud presence would want to know what Cloud services incurred the majority of its cloud expenditure. It is a common trend that only huge bills demand everyone’s attention whereas the average bills are overlooked in business.

It is therefore essential to know how much one is spending on cloud services as well as the distribution of charges. The table below lists the various cloud services and their pricing trends, which would help the company have a bird’s-eye view of its Cloud expenditure.

“It is a common trend that only huge bills demand everyone’s attention whereas the average bills are overlooked in business.”

Typically, an organization uses multiple Cloud accounts from a single Cloud Provider to meet specific requirements of various departments, local offices, or subsidiaries. The consolidated bill will include the total spend on each AWS account. Below is a representation of a sample Cloud bill with percentage spent on each of the accounts:

 

Eventually, an organization may adopt a multi-cloud strategy to avoid vendor lock-in, to utilize specific Cloud services or to implement Disaster Recovery. The consolidated bill in this case will include the incumbent costs for each Cloud Provider as shown below:


The bills educate the Corporate Finance department about the total expense on their Cloud Infrastructure. A review of the bill and charge-back are ensured in every organization. However, only a deeper analysis of the bill would show where precisely the organization has been spending on, in Cloud. The next section deals with the various aspects of itemizing and analyzing the bill and the tools that help in the process.

 

Cloud Finance Analysis

This section focusses on the different ways an organization can view fine-grained details about their Cloud expenditure.

Certain businesses may find it important to identify the Cloud cost distribution across different departments and geographies, while other organizations may need to report the costs by different Cloud resources and by the applications that run on Cloud. Companies that run their workload on multiple Cloud platforms may find it relevant to view cloud costs based on Cloud service provider. A combination of these cross-sections can also make a useful view. A daily, weekly, monthly or yearly trend can be extrapolated to make projections of usage and costs into the future.

The various tools available in the industry for Cloud reporting can range from the gamut of reporting possibilities exposed by the Cloud Provider themselves, to offerings from third-party vendors as shown in the figure 4. As a business, it is important to evaluate which tools give you the reports that best suit your portfolio.

Certain Cloud reports are very useful from the viewpoint of the CFO of a business who wants to make sure that the Return on Investment (ROI) of the company’s Cloud Investments is in line with the projections. Third-party tools, such as CloudHealth and Cloudability, offer a great number of ways to pull out many such reports.

Once we know where the wastage in Cloud resides through reporting and analysis, the next step is to take remedial measures to bring down those costs, and that is the focus of the next section.

 

Cloud Cost Control Measures

Knowing where the Cloud costs are spent is just a beginning of governing on Cloud. What ultimately has to happen is to drive in remedial actions on the information gathered. In this section, we propose various solutions to help stay money-wise on Cloud. Below are few key control measures:

 

1. Tuning the Architecture to save costs

Architecting your infrastructure and applications in Cloud is an iterative process, wherein you apply a set of principles and paradigms to your design that would make it flexible and conducive to expansion. Apart from flexibility, cost is another factor that must be considered in your design.

Cloud users have a tendency to use as many resources as possible and to employ as much high-end resources as available, for a given need. However, a mindset of planning the architecture ahead of time and avoiding overprovisioning has to be followed instead. Employing auto-scaling is a great way to avoid overprovisioning.

Use of services that allow the architecture to be loosely coupled (example, SQS service in AWS) and use of dedicated services for certain use-cases (such as Glacier service in AWS for archival, AWS Direct Connect to save on Data Transfer costs, etc.) should be encouraged. Being informed about cost cuts offered by the Cloud Providers would help one redesign the architecture to be cost optimal.

The following list offers a few architectural insights towards saving on your Cloud costs:

 

2. Choosing the right Cloud Provider and Services

It is essential that a detailed Requirement Analysis be performed before determining the right Cloud vendor/partner. The key parameters captured during in-depth application assessment and infrastructure discovery can help in determining which particular target Cloud Platform is best suited, from cost, compliance and technology point of view.

The following are some of the points that ensure cost-effectiveness in the choice of Cloud Provider/Services:

3. Shut down unused services periodically

The majority of costs in Cloud are incurred by running instances or servers and their associated data storage. Many a times, development teams spin up servers for testing and leave them running for a long time. Same is the case with other expensive services that the team might have used for evaluation or Proof of Concept projects. Hence, it is important to periodically monitor or audit the Cloud Inventory in order to save money for the IT Department of a company.

Follow some of the points below to save on costs:

 

4 Tagging Services Periodically

From the cost perspective, the relevance of tagging the services cannot be stressed enough. First, the cost reports make more sense and add value when the resources are tagged, as one can enlist user-defined tags (such as Application, Stack and Role Tags). With this information in place, one can know which application is using which service and where exactly the costs were spent. Second, it would be easier to determine which servers to downsize when we have more information handy.

Listed below are few relevant user-defined tags:

 

5. Sizing the resources

When a user has over-provisioned a deployment, it becomes necessary to detect low utilization and adjust the machine sizes. The following set of guidelines are suggested to facilitate this process:

               

6. Choosing a Cost-Effective Pricing Scheme

Most of the Cloud Providers offer a range of pricing schemes to save on Cloud. One has to tap the right pricing scheme depending on one’s usage and commitment. Choosing the right pricing scheme, however, requires experience. Tools such as CloudHealth, AWS Trusted Advisor, Cloudability3, etc. make recommendations to change the pricing plan of existing resources based on usage patterns.

The following list shows the different usage patterns and their corresponding pricing schemes:

 

 

7. Having good Cloud Governance measures

According to IDC Technology Spotlight, more mature cloud architectures depend on policy-based automation and self-service to rapidly provision new services and ensure compliance with corporate standards and service-level agreements (SLAs).

 

According to IDC Technology Spotlight, more mature cloud architectures depend on policy-based automation and self-service to rapidly provision new services and ensure compliance with corporate standards and service-level agreements (SLAs).

The IT departments feel that because of automation and self-service, the control is moving away from their hands to those who are actually provisioning the cloud services. This phenomena has led to a new term called “Shadow IT”.

Provisioning of services in Public Cloud directly impacts the total costs of services. The IT team should be in loop when decisions are made for usage of cloud resources and/or services. The Cloud Governance framework should also define the maximum spend capacity for all departments to keep costs under control. Furthermore, only legitimate users should be allowed to provision services for approved business cases.

The Cloud Governance framework should provide an answer as to who is accountable for safeguarding confidential or sensitive information stored in Cloud. There should be a clearly defined encryption and multi-factor authentication standards for all employees, contractors, and third party vendors when they are accessing sensitive data in the Cloud. Appropriate trainings could be imparted to the users who would be accessing sensitive information in Cloud.

One distinct advantage of using Public Cloud is that most of the cloud services often come with advanced monitoring, management, and tagging features. An effective Cloud Governance framework can be built by utilizing the key features of cloud services and by adhering to the published vendor guidelines.

An organization can also utilize third party CMP tools which provide multi-cloud governance modules. These modules can further be utilized for automated Cloud Governance across multiple cloud platforms. Detailed information related to cloud discovery, cloud utilization, cloud spend, cloud security, and application resiliency can be obtained easily for building the governance model.

 

Case Studies

Case Study 1 – Cloud Reporting and Analysis to save on Cloud costs:

After migrating their workload to AWS cloud, one of our customers realized that it was a humongous effort to manage their cloud costs over a period of time. UST Global, being a Cloud technology consultant, helped them get a hold of their expenses by introducing them to the various tools for mining and reporting of their Cloud spend for various time intervals. Business insights derived from these reports aided their stakeholders to take corrective actions.

AWS billing and cost management service was configured to get the detailed billing reports and cost allocation reports. AWS’s Cost Explorer was utilized to filter Consolidated Costs based on specific parameters. Also, various utilization reports provided by AWS for on-demand and reserved instances were collected to get information on purchase history and instance usage for the linked accounts.

Data mining from these reports and tools identified the exact areas like services, applications, departments, etc. which were incurring huge cloud costs in a recurring pattern. From these valuable insights, corrective actions like shutting down the instances or downsizing the instances were taken. Similarly, other third-party tools were showcased, that interfaced with the Cloud Provider’s API, for collecting useful information about the instance usage and trends. With this data, it can calculate savings and also make recommendations for changing pricing schemes.

 

A variety of automated reports were designed and implemented, that were run as weekly scheduled jobs using server-less Lambda functions. These reports were mailed to the decision-makers so that the customer could analyze their Cloud Inventory that could be cleaned up at regular intervals. Along with this, effective Cloud governance policies were introduced and the process of tagging services periodically was automated so that their reports were more informative.

 

Case Study 2 – Using various cost control measures to save on Cloud costs:

A customer with large workloads on AWS Cloud consulted UST for optimizing their operational costs. When examining their infrastructure, it was found that there were several instances that were underutilized.

The following recommendations were presented as the methodology used to tune the EC2 instances to save money.

For example, C4 instance type was used for many of their NAT servers and VPN servers. The requirement was for enhanced networking capability instances and not for high computational needs. Therefore, an M4 instance type was a perfect fit, which reduced the overall cost considerably. These recommendations coupled with decisions to shut down servers after consulting the users of the instances saved 35% of Cloud Costs per month, bringing in great revenue to the business.

The customer had not planned well on selecting the right pricing schemes for their business’s applications hosted on Cloud. After analysis, we recommended many pricing scheme switches for their applications which is consolidated into the figure 14:

The purchase plan switches helped them save more than 40% of their running costs. Accounts which were experimenting their workloads in Cloud or running development environment were identified. AWS free tier options for such accounts were suggested, which they could avail by unlinking those accounts from corporate consolidated

AWS billing account. We also defined a process to revisit their cloud architecture every 6 months to identify new workloads which are feasible for moving to economic pricing schemes.

 

Case Study 3 – Tuning Architecture to save on Cloud costs:

Here is another case study where redesigning the cloud architecture with enhanced cloud services saved costs along with increase in application robustness for a customer in AWS Cloud.

The customer’s monolithic web application was unwieldy, expensive and rigid from the design point-of-view. Their application had a time-consuming backend workload. The Web and App tiers were hosted on high capacity servers and since they had spiky demands, multiple servers were maintained.

We suggested a decoupled architecture for the application applying a scalable cloud design pattern. The new provider managed cloud services which are highly scalable and robust were used to replace some application tiers. The proposed architecture saved about 45% of their costs compared to the previous monolithic architecture they were running on. The figure 15 depicts the architecture we designed for their system:

 

Why UST Global?

UST Global has established a track record of fostering an innovation culture and providing that as added value and paid service for our clients, which has been validated by industry analysts from Forrester and Gartner. We achieve this by Centers of Excellence. We are committed to helping clients achieve business objectives by leveraging the latest in technology and innovation. We enable our associates to be proactive, responsive and flexible in our approach to business. This nurtures the spirit of innovation in each and every associate. Our Centers-of-Excellence (CoEs) are made up of experienced industry experts that support our major verticals in a consulting model. CoEs are established for discovering and applying industry-leading knowledge and competency to improve the quality and efficiency of our service delivery. Our Digital Transformation and Cloud Services CoE provides consultation for several cloud projects within and outside UST. We currently have design and deployment expertise in public, private, and hybrid cloud models for various business verticals.

UST Global's business philosophy of "fewer CLIENTS, more ATTENTION" is a distinguishing factor enabling us to provide unsurpassed business value to our client, by way of a highly tailored and customized solution. Our focus is on client mindshare, not market share, and clients continually rate us ahead of our competitors in attributes such as speed of response, flexibility, innovation, and 'Commitment-beyond-Contract'. Customers will experience this first hand in every interaction with the UST Global team. UST Global follows a client-centric engagement model wherein we endeavor to develop partnership with clients through focused senior executive oversight and management.

UST Global is a partner with AWS, Azure, Google Cloud, and VMWare. UST has experience helping customers migrate large workloads to Cloud, application assessment, infrastructure discovery, cloud security, and cloud governance. Our proprietary CMP tool FogPanel is a single pane of glass management solution for a multi-cloud environment.