The Value of Building a Multi-Cloud Strategy
September 22, 2016
The latest IaaS Gartner quadrant report put AWS, Azure and GCP as the public cloud market leaders, but which of them best support the specific needs of your enterprise? For example, how’s the durability of S3 object storage on AWS compared to Azure blob storage, and how’s the performance of AWS Redshift compared to Google Bigquery?
How can you build the most value from the cloud? What will help you to gain bargaining power for better workload mobility and prevent vendor lock in? How can you capitalize on the cloud’s distinct advantages – for cost savings, risk reduction, better business agility and IT system flexibility?
Upon implementing a multi-cloud strategy, it’s often difficult for enterprises to mix and match different Infrastructure as a Service (IaaS) offerings. So, let’s consider the value and challenges associated with running multi-IaaS workloads in AWS alongside other cloud providers like Azure and Google.
Implementation Challenges Ahead
The process of migrating applications and data to a single cloud service provider is challenging enough, but deploying and running multiple clouds is even more complex. Multi-cloud implementation requires in-depth knowledge of the cloud services as well as the ability to compare services across clouds.
In addition, there’s very little standardization. Each service provider has unique technology and capabilities – from the basic building blocks of the main subsystems (i.e., compute, storage and network) to its APIs. Therefore, best practice is to keep the workload independent and decoupled from the underlying IT infrastructure which can allow mobility between different environments. However, this decoupling is not always feasible, especially for enterprise legacy workloads.
Working with multi-cloud also requires more strategic expertise and a broader technical skillset. IT staff must understand how to manage all the aspects of each cloud vendor environment, including performance, security and availability.
Why Do It?
Despite the challenges, CIO’s and IT teams look to build a multi-cloud strategy because it can offer substantial cost and agility benefits. Here’s what they strive to achieve:
1. Gain autonomy by reducing lock-in
One of the strongest advantages for on-premises environments is autonomy and control, which have traditionally been compromised in a move to the cloud. Therefore, while the advantages of the cloud may make many enterprises willing to relinquish this control initially, lock-in becomes an issue later. For any IT system, relying on a single vendor is never a good idea, and the public cloud is no different.
In addition, as technologies evolve and the enterprise’s IT environment needs to change, a multi-cloud approach enables flexibility for its infrastructure to become more robust. During contract negotiations, this flexibility also provides leverage – very relevant for enterprises and managed service providers (MSPs) with large cloud footprints. When your cloud no longer depends on one vendor and can easily be moved elsewhere, there’s more freedom in decision-making.
2. Reduce operating cost with better price/performance
Cloud service providers offer a wide range of pricing models and packages, with granularity to bill by the hour or even the minute (e.g., AWS vs. MS Azure). When comparing alternatives, significant differences may exist that can have a direct impact on the IT environment and its operating cost. In some cases, IT management may want to choose the cheapest available option, while in others,IT leaders may prefer to spend more for a specific deployment to achieve special capabilities or better performance.
When comparing Infrastructure as a Service offerings across clouds, mix and match may be the best decision. For example,choose Google for its fast instance launch. Pick AWS for its mature and comprehensive instance portfolio. Add Azure due to better custom packages in case your organization is under the Microsoft Enterprise Agreement (EA).
3. Extend capabilities with more flexibility
Gartner recently published their interesting report “Critical Capabilities for Public Cloud Infrastructure as a Service (IaaS), Worldwide” which shows the range of features and capabilities between leading cloud vendors. Below for example a chart from the report with scores per cloud vendor’s capabilities in support of enterprise “critical capabilities.”
For example, AWS is well suited to organizations with a large number of users, for large-scale data processing and for meeting most security and regulatory compliance requirements. However, for customers with existing Microsoft deployments, Azure is a particularly good fit and comes close to AWS on capabilities like scaling and provisioning VMs, large-scale data processing and storage, and enterprise integration features, such as data replication. For single VM availability and reliability, the computing resilience of CenturyLink, Virtustream or Google might serve best.
With multi-cloud flexibility, one cloud vendor might offer a more desirable feature set that would justify switching. Some features, like Database as a Service, security controls, compliance or high availability (HA) architecture, might not be supported by all vendors. Therefore, shopping around for capabilities that may benefit your enterprise, while continuing to lower resource costs becomes a key benefit for more flexibility and extensibility.
A multi-cloud approach can deliver undeniable benefits. It’s best to get your enterprise on board sooner than later and identify the challenges now. Invest in IT training to understand how and when to extend your datacenter to the cloud, which will include establishing workload mobility, controlling data residency, and understanding backup and DR implications.
How? The tools and platforms exist today to address most of these challenges by facilitating workload mobility and creating a unified management platform that can span your on-premises data center and your cloud instances. Don’t get caught in doing only what you know and get stuck with on-premises or single cloud opportunities — spend the effort and training required to master the necessary cloud skills. It’s the inevitable future.