How to optimize your cloud usage

A 2019 Technalysis Research study found that medium to large businesses on average use 3.1 Cloud Service Providers with 3.4 cloud workloads per provider; over 85% of these businesses were also running a private cloud. Additional data was offered regarding the 13 most popular workloads deployed across public, private and hybrid clouds, from databases (72%) down to security (46%.) Other categories included Analytics, IOT, Software Development and Backup/Disaster Recovery. All of this data reminds us that businesses need to review their own cloud usage and assure an optimized plan exists that aligns cloud and storage models with workload performance requirements and risk profiles. Creating or revising a cloud usage plan has significant impacts beyond performance and security; key among those is the economics of cloud usage.

Cloud-onomics Has Significant Impacts on the Business

When a business doesn’t have an optimized plan for cloud usage, there are many potential side effects. A survey by 451 Research and Apptio offers that inferior cloud management has a negative impact on innovation, provides a lower quality of service, promotes sprawl and underutilization of resources, and leads to higher costs. Some estimates put the 2020 “wasted cloud spend” at $17.6B, made up of categories that are largely correctable with a properly optimized cloud plan. The categories include:  nonproduction (44%), oversized (40%), overspend (50%), and idle workloads (76%).

Segmenting Workloads

Reviewing and segmenting workloads is a key step in optimizing a cloud usage plan. Following is a high-level overview of workload considerations for public, private and hybrid cloud services:

• Business critical workloads like CRM/ERP, Database, BI, Analytics, etc. are those that are most demanding of compute, storage and security services. Further up the critical workload stack, demanding High Availability (HA) service levels, are mission critical applications; think transportation grids, 911/emergency and healthcare records. Business and mission critical workloads typically require the performance, security, availability, SLA, and compliance regulations associated with private and hybrid cloud. While the expense of hybrid or private cloud vs public cloud is a factor in the cloud usage plan, it is important to consider that HA levels of service are not typically architected using public cloud. When you consider the 24/7/365 nature of business and mission critical workloads, the metered costs of public cloud services can grow rapidly exceeding private cloud costs. Managed private cloud takes workload protection to a higher level featuring 24/7/365 first responder services and expertise that acts as an integrated extension of your security and IT staff.
• Public cloud is a multi-tenant, highly dynamic and expandable environment that offers shared resources with data and applications secure and invisible to other tenants. In the optimized cloud plan, examples of workloads segmented to the public cloud are development platforms and testing, large data processing projects, public information websites, some SaaS/CRM applications, global access services and training, to name a few.
• Similar to segmenting workloads and selecting the appropriate level of cloud service, backup and disaster recovery (DRO) need to be considered by workload. Two basic definitions associated with backup and DRO; Recovery Time Objective (RTO) defines the maximum allowable downtime for an application or process. Recovery Point Objective (RPO) defines the maximum allowable data loss an application or process can undergo. (e.g. time between backups.) Both of these objectives measure recovery vs potential for damage to the business. Typically for DRO services, the shorter the time interval required to achieve the designated RTO and/or RPO, the higher the cost. In order of ascending costs, the levels of available DRO are best defined as:
  • Tier 3, Data Centric. Used for S3 compliant data storage, RTO/RPO measured in hours to days.
  • Tier 2, Business Important, RTO/RPO measured in hours.
  • Tier 1, Business Critical, RTO/RPO measured in minutes. Tier 1 is typically a fully managed service (DRaaS) to provide the most effective levels of critical backup.

The Value of SLAs Considered in Cloud Usage Plans

Commensurate to the value of the workload and the cloud service the workload is segmented into, SLAs are a key component of the cloud usage plan. As previously noted, the workloads that require private or hybrid cloud should also have more stringent SLAs for availability, RTO, RPO, security and compliance. In either service, public or private cloud, higher SLAs normally have a higher cost which should be justified against the value of the data or workload protected. All cloud service SLAs should consider:

• Availability, usage and uptime
• Performance benchmarks
• Service time and maintenance
• Support/MTTR metrics
• Resource parameters and limits

Developing or revising a cloud usage plan should include segmenting of workloads, optimizing those workloads to the appropriate level of cloud services and selecting commensurate levels of DRO and business continuity. These choices involve complex decisions, complex decisions that the industry’s best expertise can assist with. If your organization, large or small, is preparing to develop or optimize your cloud usage plan, we can help! Otava’s fully compliant public and private cloud services answers your workload requirements and Otava disaster recovery as a service, powered by Zerto, helps you stay up and running in the face of cyber-attacks, natural disasters and system failures. When you sign up for 12 months of Disaster Recovery as a Service, your first month is free! Contact us to learn more.

Related Articles

The Exploding Use of Clouds; Which Cloud is Best? The numbers reported for the expanding use of all cloud services today is staggering. Gartner shows public cloud spend exploding from $91B in 2011 to $227B in 2019. 451 Research forecasted 90% of organizations using some type of cloud service…(read more)

Do You Have the Right Levels of DR and backup for each workload?  Deploying Disaster Recovery or Disaster Recovery as a service on a per workload basis would be a fairly straight forward topic if only reviewing the technical aspects of implementation. (Read more)

Hybrid Cloud What It Is, Why It Matters: How is hybrid cloud defined, and why should an organization consider hybrid cloud? The standardization of definitions and the use of cloud by 94% of enterprises has made hybrid cloud an essential service category. (Read more)