Multi-Cloud Object Storage and Data Mobilityv1.0

1. Summary

Cloud object storage is becoming increasingly relevant in enterprise due to its intrinsic capabilities: global data distribution, data durability, unmatched scalability into the exabytes, and overall cost-effectiveness. Solutions are easy to adopt, as they are offered by public cloud providers or on-premises. Data is accessed through standard HTTP-based protocols and can be written and retrieved anywhere from any device. However, making data accessible to all remote and local applications and users securely, at a low latency across multiple clouds, is still very challenging. This leads to a rise of application complexity, higher costs, and limited freedom in the execution of a multi-cloud strategy. With the rise of edge computing and IoT applications, data is now created in several places and then consolidated in large repositories. These repositories must be accessible to an increasing number of applications and workloads, ranging from real-time web apps to big data analytics. This poses increasing challenges, making it more complex to respond quickly and adequately to changing business needs.

This report clarifies the challenges posed by multi-cloud storage, examines how to realize a winning multi-cloud storage strategy, and identifies the right tools to succeed.

We will analyze several aspects of multi-cloud storage including:

• Why to utilize a multi-cloud storage strategy
• Data security, protection, and availability challenges
• Policy based management and service levels
• Automated tiering
• Data movement across clouds
• Benefits of a consolidated view of data and services
• Common back-end and front-end protocols
• Key players

Key findings include:

• End users want to avoid lock-ins, the flexibility made possible by a multi-cloud strategy allows them to take advantage of the best services from each single provider; but, to be efficient, data should be as close as possible to compute resources.
• Although end users are asking more from multi-cloud data management solutions, most vendors still lag behind, with products that offer only a subset of the required features. The market is evolving quickly though, the tools to manage data across clouds; move it seamlessly; or provide a transparent, single domain view are surfacing. We classified available tools in three categories to reflect their capabilities: Object store with tiering functionalities, hybrid object stores, and multi-cloud controllers.