How a Multi-Cloud Strategy Optimizes Cloud Workloads

A Multi-cloud strategy leverages multiple cloud providers to optimize workloads and business outcomes.


No-Size-Fits-All! An Application-Down Approach for Your Cloud Transformation

Big-box discount stores are a go-to store for inexpensive shopping. Customers don’t get or expect personal attention; they’re there for big selection and low cost. But when those customers need a fine suit for that make-it-or-break-it meeting, they don’t go to there. They go to a higher end store for high quality and personal attention.

A similar choice exists between hyperscaled public clouds that try to offer everything under the sun, and specialty Cloud Service Providers who tailor their cloud environments for specific workloads and highly customized customer support.

Choice is the meat of the multi-cloud strategy. Multi-cloud is an approach where cloud customers do not need to choose between single clouds, but instead choose multiple clouds to optimize workloads and business outcomes. The multi-cloud mix may include private clouds, hyperscaled public clouds, customized CSPs, and hybrid clouds.

Cloud Definitions

Before we dive in to a multi-cloud strategy let’s agree on some definitions, especially since customers, vendors and analysts throw around cloud terms with somewhat wild abandon. These definitions are generally well-accepted in the industry.

·  A private cloud is a sole tenant cloud infrastructure with virtualization, service orchestration and dynamic provisioning. Most private cloud owners locate the infrastructure on-premise. Although the organization may contract IT companies to help manage their virtualized private cloud, the business is its sole owner and tenant.

·  Public clouds are multi-tenant computing environments that host a variety of different workloads and environments. This large category includes hyperscaled public clouds and customized Cloud Service Providers, or CSPs.

·  Hyperscaled public clouds offer a wide variety of services including SaaS, PaaS, IaaS, application data storage, and long-term secondary storage. They are massively scaled and geo-located in many locations throughout the world. The Big Three are AWS, Azure and Google Cloud.

·  Customized CSPs customize their offerings by workload type, vertical industry or geographical location. The strongest competitors in this category offer customized workload and customer support that is difficult to get from the hyperscaled public cloud.  Backup and disaster recovery (BDR) provider KeepItSafe and PaaS providers DigitalOcean and Virtustream are leading examples.

·  Hybrid clouds orchestrate services running on multiple clouds, usually private clouds that cloud-burst or tier to a public cloud. Another common usage is a single-vendor cloud that extends on-premise customer deployments to the vendor-owned cloud. Of these, Oracle, SAP and IBM clouds rival the hyperscaled public clouds in yearly revenue.  

·  Multi-clouds can include all the above cloud types. Since it does not orchestrate between cloud services, multi-cloud is less a technology and more of a strategy. This will probably evolve as developers improve orchestration and workload migration tools between clouds, but today these tools are in the nascent development stage. For many companies it’s enough to know that contracting with multiple clouds is good business.

Why Multi-Cloud?

To adapt an old saying to the hyperscaled public clouds: “Jack of most trades, master of few.” Some specific services are unmatched, such as massively sized IaaS on Azure or high-performance application hosting like Salesforce on AWS.

But simply because the hyperscaled public cloud offers a lot of services, does not mean they are optimized for all those services. Then why do some businesses still place all their cloud services on the hyperscaled public cloud? There are several reasons including familiarity, and perceived low cost and simplicity. (Note “perceived.”)

·  Familiarity. Organizations frequently use a single usage case to pilot the move to the cloud, such as Office 365. This is fine, but over time they contract for additional services on the same cloud simply because it’s familiar, not because it’s the optimal environment for all workloads.

·  Perceived low cost. Hyperscaled public clouds often have low entry price points. However, lower prices are not a given. The more services the customer adds the more expensive it gets and going over data storage thresholds can be very expensive.

·  Perceived simplicity. IT often prefers to get a single bill from one provider rather than several bills from different providers. However, hyperscale public cloud billing can be complicated when billing for multiple services and having a single vendor for multiple services is no guarantee of simplicity.

·  Support. Customer support is a major concern for cloud customers. But pouring resources into customer support is a money-loser for the hyperscaled public clouds. Their profit model hinges on customer DIY, and the less time and resources they spend on customer support the better for them. Customer support becomes a major differentiator for CSPs who put resources into it. And some CSPs differentiate themselves even more by customizing managed and monitored services to their customers. Customers get top level support and the services they want, and the CSP differentiates itself even more from the hyperscaled public cloud.

The Customized Cloud: Choosing the Right Providers for Business Outcomes

Ultimately there is no clear-cut compelling reason to go with a single cloud provider. Instead of going with a one cloud fits all, which it does not, companies are better off with the ability to process workloads on clouds that specialize in that type of service.

There is no need to give up the hyperscaled public cloud services that work well for you. Keep those services but consider adding private and hybrid clouds as well. And strongly consider reaching out to customized CSPs to optimize workloads.

These CSPs approach cloud services from the standpoint of the best of private and public clouds: the high performance and rapid recovery of the private cloud with the dynamic scalability of the public cloud. In addition, leading CSPs also offer optimal workload environments, highly engaged customer service, customized SLAs, flexibility, and cost savings.

Additional benefits include services that the hyperscaled public clouds do not offer because it does not make economic sense for them. These are services like tailored retention policies or the ability to archive data for eDiscovery and compliance. Another benefit is granular security levels for SOC 2, HIPAA, FINRA, FIPS 140-2, ISO and PCI compliance and ISO 27001 certification.

Benefits of Customized CSPs in Your Multi-Cloud Strategy

·      Optimize workloads. This is the major driver for going to a multi-cloud model. BDR is one primary usage case. Hyperscaled public clouds protect customer data, offer failover for disaster recovery and cost-effectively store cool/cold data. However, they are not in the data protection business across multiple applications and clouds. For that level of data protection, customers go with customized CSPs like the KeepItSafe cloud. PaaS for application development/DevOps is another primary usage case for optimizing workloads. The hyperscaled public clouds are popular PaaS providers but focus their services on developers instead of operations. Customized CSPs can do a much better job of serving both customers with rich policy management, application integration, high availability, DR and customized SLAs.

·      Save money. Multi-cloud strategies enable businesses to choose the best pricing for the best business outcomes. They are not stuck paying for a less-than-stellar cloud backup service, or an expensive IaaS offering just because they want to stay with a single cloud provider.

·      Flexible goals. Different applications have different requirements. For example, when you host your ecommerce application in the cloud, then performance and availability are extremely important. When you store sensitive business data, searchability and governance take the top priority. And flexibility is the earmark of a good PaaS choice, while extreme scalability is primary feature for IaaS. Multi-cloud gives customers the flexible choices they need.

·      Avoid vendor lock-in. No one blames cloud vendors for hanging on to their customers. What customers resent is being locked in to a cloud provider because moving data equals time, money, and potential loss. Another example of lock-in is being tied to a specific software platform because it is the only product the cloud provider offers.  The best CSPs will offer a choice of software platforms and efficient data migration between them for their customers.

·      Customize SLAs. Hyperscaled public clouds are reluctant to customize customer SLAs. It costs them money and resources, and many customers push back when told they need to foot the bill. Customized CSPs use tailored SLAs as a competitive differentiator by working closely with their customers to devise best SLAs and outcomes.

Making the Leap

Hyperscaled public clouds are not the only game in town. They never were. They manage some workloads and environments just fine. But instead of sticking to a single model for all your cloud computing needs, consider adding customized CSPs to your cloud portfolio.

This multi-cloud strategy lets you optimize your cloud investment and enables you to achieve the best business outcomes that your organization wants and needs.

Tags: cloud computing, AWS, IBM, Azure, Salesforce, cloud workloads, multi-cloud

0 Comments (click to add your comment)
Comment and Contribute


(Maximum characters: 1200). You have characters left.



IT Management Daily
Don't miss an article. Subscribe to our newsletter below.