Nmedia - Fotolia


Closely manage your hybrid cloud approach

Hybrid clouds are a fit for many enterprises, albeit convoluted to set up. IT pros must identify the right cloud providers for workloads and develop strategies to overcome hurdles.

Pursuing a hybrid cloud is not for the faint of heart. In the short run, the best hybrid cloud approach may be to identify applications that can cleanly run in different environments for long durations. Simply moving workloads around is just the tip of the iceberg. It's important to assess the management challenges of trying to orchestrate applications spanning different clouds. Enterprises must also recognize factors that could affect hybrid cloud performance, including application selection, data migration, latency issues and disaster recovery tactics.

As enterprises contemplate hybrid cloud strategies that span Amazon Web Services (AWS) and other cloud infrastructures, they need to choose where to stage data, services and applications. The enterprise will have to examine the business value of this kind of migration upfront to determine which workloads should be migrated. There also needs to be some thought put into the management challenges of workload migration and application performance for apps that span a cloud infrastructure.

Using VMware infrastructure makes a lot of sense for migrating legacy applications into a cloud environment. And OpenStack is gaining ground as another approach for deploying applications on a consistent infrastructure that can be migrated on premises or to private clouds.

One factor to consider when moving to hybrid cloud is how to properly stage data to minimize the migration of large data streams required for analytics. Latency can also play a factor, as round-trip cycle times can increase when services cross cloud boundaries.

And while hybrid cloud backup is a relatively straightforward way of providing disaster recovery, creating an infrastructure for fluidly and reliably moving workloads is more challenging. Another key challenge lies in simplifying API access to applications spanning a cloud infrastructure. Some important considerations of creating applications that move across a cloud infrastructure include application clustering, operating system (OS) versions, support for shared file systems, IP address allocation, load balancing and database I/O requirements. These challenges can arise because of subtle differences that exist in implementing similar features across different cloud platforms.

Finally, enterprises should consider how each cloud provider in the runner could meet their governance, risk and compliance requirements.  

The rise of AWS alternatives

Amazon Web Services quickly evolved into the largest cloud provider because of its early start in the industry, flexibility, and rich ecosystem of development and deployment tools. It is also one of the leading candidates for new application development among enterprises starting from scratch. But organizations may consider VMware as a better alternative for legacy workloads, which allows for a migration free of refactoring and rewriting. In addition, OpenStack's open source nature and support from a number of leading enterprise software vendors make it easy to move across different cloud providers' infrastructure.

VMware was one of the first companies to provide a cloud-like operation system for enterprises, because it developed an infrastructure for quickly spinning up existing application workloads onto a platform that mirrored an existing OS infrastructure. At the same time, it has been a latecomer in rolling out a public cloud service. VMware vCloud Air is suited for mixed Linux and Windows workloads; it also uses the same vCloud infrastructure for internal and external cloud environments. This keeps management simple.

A number of leading enterprise software vendors, including IBM, Google, Dell, Red Hat and Hewlett Packard Enterprise, support OpenStack. But there are differences in OpenStack implementations, and challenges in automating configuration and cloud bursting. OpenStack implementations are built on top of an array of modules for compute, storage and networking services.

Migrating simple workloads between clouds

One-off migrations of simple application workloads between VMware and AWS are relatively straightforward. It becomes more challenging for complex applications, where there is a need to constantly move workloads across cloud boundaries or when moving between AWS and OpenStack.

The AWS infrastructure is focused on the public cloud, with limited off-site cloud management. Implementing applications that harmonize with AWS requires implementing a direct connecting service to bridge an enterprise's data center with virtualized private cloud resources.

Tools such as AWS Direct Connect can help simplify the process of migrating workloads to AWS infrastructure and back out to VMware, but are lacking when it comes to support for OpenStack. Simple VM/AMI migration works well for moving self-contained workloads, but it can present a number of managed and application integration challenges if the enterprise wants a truly fluid infrastructure across cloud platforms.

Some level of refactoring is required for all but the simplest applications. Translating an individual VM to function across a cloud infrastructure is fine for simple applications. But this hybrid cloud approach is sorely lacking for the kind of applications built on top of multiple services.

Managing hybrid workloads

The best strategy when spanning multiple clouds is to use third-party hybrid cloud management and workload migration tools from companies such as RightScale, Platform 9 and Mirantis. These tools can help address the hurdles and complexities related to building a fluid hybrid cloud environment.

Leading OpenStack providers are starting to embrace AWS as well, which promises to help provide a comprehensive service and support for multiple cloud infrastructures. For example, Rackspace recently announced a major initiative to support AWS, which complements its existing suite of OpenStack services and support.

Supporting APIs across a hybrid cloud

Another challenge is that many cloud management platforms support AWS-specific APIs directly. Organizations that want to combine internal resources with AWS resources may need multiple APIs, which create additional overhead and development complexity.

One hybrid cloud approach is to take advantage of EMC's open source tools, which provide AWS API compatibility for OpenStack infrastructure. Another hybrid cloud approach is to use translation services, which provide a broker for AWS offerings on top of OpenStack and VMware infrastructure. Cisco's Intercloud Fabric provides a common API across multiple public cloud services. EMC's Enterprise Hybrid Cloud simplifies API access to applications running on VMware, OpenStack and AWS clouds.

Next Steps

Hybrid cloud management is no easy task

AWS offers hybrid cloud savings

Hybrid cloud with AWS and OpenStack is a balancing act

Dig Deeper on Amazon EC2 (Elastic Compute Cloud) management