blog

Making Sense of Hybrid IT

Andrew McBathYour customers are more aware than ever of the critical role the IT infrastructure plays in delivering competitive advantage. And as they aim to optimize performance, cost-efficiency and security for a growing range of applications and workloads, interest in hybrid environments is growing quickly. Recognizing this opportunity, service providers are rushing to define what “hybrid” means and basing offerings on their own interpretations. 

Amid that noise, several definitions for the term stand out, but each has limitations.

For example, many experts, including the National Institute of Standards and Technology (NIST), focus exclusively on the cloud when describing hybrid environments. NIST defines a “hybrid cloud” as a combination of public, private and community clouds “bound together by standardized or proprietary technology that enables data and application portability (e.g., cloud bursting for load balancing between clouds).”

While this definition is sufficient when describing a combination of disparate clouds, it fails to address common scenarios where technologies enable data and applications to be easily managed and moved across both cloud and non-cloud infrastructure environments.

Hybridization is often also defined as a mixture of on-premises and hosted cloud — but this presents a number of problems. First, the importance of the demarcation line between on- and off-premises infrastructures is diminishing. Logical and physical networks and other infrastructure elements, like WAN acceleration appliances, firewalls, storage gateways and application-delivery controllers, are routinely extended across companies’ on-premises data centers and third-party sites to improve network performance, security and ease of use. This blurring of the cloud “edge” renders the physical location of an application running across environments less important. Split application architectures may be deployed in a single data center but have certain workloads reside within different hosting environments.

Still others describe hosting and cloud platforms connected via unmanaged network links as a hybrid environment. In these scenarios, the service provider runs bulk cable to a facility meet me room and connects to another cloud, hosting or colocation vendor. These links don’t typically enforce network transparency across environments and can make it difficult for users to logically provision, manage and monitor machines identically in both their legacy and interconnected, third-party cloud environments.

The need for a standardized definition of hybridization is becoming increasingly clear. Partners need to be involved so they help customers make sense of the growing number of approaches.

Hybrid Is as Hybrid Does

While many of the technologies that link disparate infrastructure environments still have limitations and room to mature, there are certain elements to look for when assessing a service provider’s ability to deploy fully functional hybridized solutions today. These include:

  • Flexible service offerings that can meet a full range of application requirements, spanning public, private and bare-metal cloud, as well as managed hosting and colocation services. The location of the “execution venue” – on-premises or off – isn’t significant. What is important is the ability to easily leverage the best service or execution venue for each application.
  • A unified network fabric for deploying compute, storage and networking services across environments. Hybrid network topologies must enable workload mobility but face the challenge of linking abstracted cloud environments with legacy infrastructures that are still closely tied to underlying physical devices. Providers using extensible Layer 2 domains within and across data centers and virtualization-aware networking equipment, along with newer technologies such as software-defined networking platforms that overlay existing infrastructure, can address this. Among other benefits, a unified network can allow workloads operating in different hosting environments to share the same network infrastructure elements, such as firewalls and load balancers, while maintaining persistent connections to quickly move workloads across hosting environments.
  • A single-pane-of-glass interface lets customers or partners easily manage an entire IT infrastructure, including migrating workloads, monitoring SLAs and providing for sufficient network isolation and security – across servers, storage or other resources – through a centralized portal. Additionally, organizations should look for APIs that offer the ability to integrate on-premises infrastructure with hybrid cloud and colocated resources in the same management interface.

A single point of contact eliminates the need to deal with multiple service providers, and a single bill allows customers to see the details of exactly which resources are used for which applications, making it easier to manage costs across the relevant groups within an organization.

By getting familiar with the characteristics that establish a provider’s ability to deploy true hybridized environments, partners can ensure they choose a solution that meets all of a customer’s application needs, both today and in the future.

Andrew McBath is Internap‘s senior director of market strategy. In this role, he analyzes customer adoption and usage trends and influences the go-to-market approach for the company’s Internet infrastructure strategies. .


Leave a comment

Your email address will not be published. Required fields are marked *

The ID is: 92105