The Multi-Cloud Advantage

Approaches to multi-cloud range from choosing different platforms for different workloads to dynamically moving workloads to take advantage of vendor specific pricing and performance. Cloud performance monitoring gives you the insight you need to make informed management decisions

MORE THAN ONE CLOUD PLATFORM IS NOW THE NORM Multi-cloud is getting a great deal of attention these days. Eight out of ten enterprises surveyed by Forrester in 2018 described their cloud strategy as being multi-cloud. At this point, the industry has not standardized on a single definition. Some enterprises consider multi-cloud the use of private and public clouds for different workloads. Other enterprises describe multi-cloud as the simultaneous use of more than one public cloud for the same workload. Whatever the approach, the goal is to match workload requirements with the platform that provides the best combination of price, performance, and manageability.

ONE SIZE DOES NOT FIT ALL

Choosing the best platform for a particular workload depends on many factors: application design, integration requirements, the location of users or data, the need for security, and the availability of management tools. Platform providers, both private and public, also differentiate themselves by offering unique features and add-on services to attract specific types of cloud users or industries. Some users may want the control that using an open cloud platform provides. Any or all of these factors can play a role in the initial platform decision. Also, the ease of acquiring cloud services has led many non-IT groups to purchase cloud services, further increasing the diversity of platforms which IT is expected to manage.

FACING TOUGH COMPETITION, CHURN, AND CONSTANT CHANGE, MANY COMPANIES TEND TO PURSUE THE SAME DIGITAL INITIATIVES AS THEIR PEERS PLAY ‘CATCH UP’, ACTING ON THE FEAR OF MISSING OUT, OR WORSE, FEARING EXTINCTION

MULTIPLE CLOUDS REDUCE RISK

In addition to pure functionality, some organizations spread a single workload across multiple cloud platforms to increase fault tolerance. While uncommon, outages and breaches do occur at even the largest cloud providers and this is an unacceptable risk in some industries. With a single platform, recoverability is limited to what one vendor can offer. A multi-cloud deployment lets cloud users shift processing to another vendor, if necessary, to recover quickly from the loss of a region or an entire provider.

Dividing processing between two vendors also provides an insurance policy against rising costs. With contracts and payment procedures in place at two providers, a company can scale up processing on the most costefficient platform as demand grows. Many organizations employ a version of this strategy as a hedge against price increases. The chief hurdle to switching vendors is dependence on platform-native tools or services. The industry is responding with cloud management platforms (CMPs) and cloud service brokers (CSBs) that separate cloud management from the underlying cloud platform, to reduce lock-in and facilitate a smoother transfer.

WORKLOAD TRANSFER CAN IMPROVE PERFORMANCE

A multi-cloud strategy also gives organizations options for improving application and network performance. Before a platform switch can be justified, the IT team must first gain visibility to performance across the distributed environment and identify the metrics they will use to making platform decisions. The next critical step is to run simulations on alternate platforms and validate their ability to meet required service level agreements. In the cloud, data transport is often outside the control of the enterprise and has a significant impact on enduser experience. Some applications are susceptible to packet delay, jitter, or loss. A critical step in multicloud management is simulating a realistic mix of complex, high-volume traffic across diverse end-points to observe and document any performance advantage. Validating the new platform before implementation will increase the chance of a successful rollout with minimal user disruption.

Maintaining and troubleshooting a distributed network is a significant cost to an organization. Some organizations already use active monitoring to continually measure performance, proactively identify bottlenecks, and test fixes. In a multi-cloud context, active performance monitoring can lead to a cost reduction and also provide information critical for cloud orchestration.

THE FUTURE OF MULTI- CLOUD

At the leading edge of the multi-cloud strategy are enterprises building an environment where potential workload transfers are automatically identified and dynamically executed. In this environment, policies and tolerances are pre-set, and key metrics are continually monitored to determine when a transfer is advantageous. An orchestration platform initiates and completes the workload transfer between platforms. This strategy is getting an assist from container technology and new, cloud-agnostic management tools designed to enable resource configurations to be used on multiple platforms and keep track of cross-cloud dependencies. Eventually, these tools will reduce the manual effort of cloud orchestration and allow organizations to build large-scale multi-cloud environments.

A truly dynamic approach to multi-cloud is not yet a reality for most enterprises. In the meantime, enterprises should establish practices and tools for monitoring and measuring performance in the cloud to build the knowledge they need for dynamic cloud management. As multi-cloud technology matures, the ability to switch platforms without significant expense or manual effort is an insurance policy against unexpected rising costs, failure to innovate, or unacceptable service and support.