The Emerging Need for Workload Acceleration

A New Level of Agility for the Virtual Enterprise

As enterprises deploy Virtual Machines (VM) in geographically diverse data centers for resiliency or load-balancing reasons, IT managers must contend with network performance issues as traffic moves from one location to another. Network latency, congestion, and capacity (i.e. <a href="http://bizibl.com/technology/what-is/bandwidth" target="_blank">bandwidth</a>), for example, all can adversely impact workload performance across a distributed enterprise. (A “workload” is defined as one or more applications running as VMs).

In the traditional IT environment, the workload’s performance depended on coordination across IT teams – application, server, storage, and networking. The result was more process, more overhead, and often delays in meeting business requirements. Server virtualization enabled IT to remove some of the barriers between compute and application teams, but now with network virtualization, and more specifically virtual network services, the barriers separating the networking team will start to come down. Application owners will be able to gain even greater visibility and control over their applications while making networking teams even more efficient.

By separating network services from the physical equipment and the underlying network infrastructure, these services can also be abstracted and orchestrated as a common pool of resources for end-users. As such, the tools for provisioning, configuration and management of those services can be radically simplified and automated. Network services, once exclusively the domain of networking specialists, can be virtualized and off-loaded to virtual administrators and application owners responsible for an application’s experience.

Virtual network services give application owners far greater control and transparency into the life cycle of their applications, improving the velocity in which they can deliver services and making networking teams more efficient. The combination of these benefits make network virtualization, and virtual network services in particular, essential components to realizing the software-defined data center.

Workload acceleration is a virtual network service enabling applications to perform more effectively across the WAN. With workload acceleration, application owners and virtual administrators select an individual workload from a management screen to accelerate its performance over distance. Workload acceleration is not just WAN optimization software, but a tightly integrated component in a software-defined data center. Workload acceleration brings WAN optimization technology to virtualization, storage and application owners in today’s enterprise.

Understanding the Software-Defined Data Center (SDDC)

The network is often one of the biggest challenges to quickly deploying new workloads. The Software-Defined Data Center (SDDC) tackles this problem by enabling IT to decouple workloads from underlying hardware, bringing unprecedented flexibility and control.

At the heart of VMware NSX, for example, is an extensible platform that enables network services to register with the VMware NSX controller, and seamlessly insert their respective capabilities into virtual networks. (See Figure 1)

This results in a number of SDDC benefits, which include:

• Operational Savings: SDDC lowers operating expenses. Network services can be packaged in a way that application owners can perform actions on their own, freeing up the networking team.
• Flexibility: SDDC creates flexibility in how the network can be used and operated. Organizations can write their own network services using standard development tools.
• Improved Uptime: By eliminating manual intervention, SDDC reduces configuration and deployment errors that can impact the network.
• Better Management: IT can use a single viewpoint and toolset to manage virtual networking, computing and storage resources.
• Planning: Better visibility into network, computing, and storage resources means IT can also plan more effectively.
• Infrastructure Savings: Separating route/switching intelligence from packet forwarding reduces hardware prices as routers and switches must compete on price-performance features.

SDDC enables IT to create a flexible, agile fabric that can be reconfigured as quickly and efficiently as the workloads that run over them.

Utilizing SDDC Principles to Accelerate Application Performance

As workloads move throughout a distributed enterprise, specific tools are required to ensure that network limitations, such as bandwidth, latency and congestion do not adversely impact application performance. Just as SDDC simplifies the actual transport of data across an enterprise, workload acceleration has emerged as a way of improving application performance across a distributed enterprise.

With workload acceleration, virtual administrators can select which workloads need acceleration as they move throughout a distributed enterprise. With a simple point-and-click, traffic is routed between the virtual switch and an acceleration engine, such as a Silver Peak data acceleration software instance. (See Figure 2)

The acceleration engine applies various real time techniques to condition the data for travel between sites. These optimization techniques might include data deduplication to maximize available network capacity (i.e. bandwidth), latency mitigation to overcome distance limitations due to chatty protocols, Quality of Service (QoS) to prioritize workloads, and loss correction techniques to minimize the ill effects of network congestion (i.e. dropped / out of order packets).

While the specific workload acceleration techniques will vary by vendor there are the following key characteristics that typically exist:

• Workload Level Granularity: The acceleration engine should have application visibility at the workload level, giving insight into the performance of all applications running throughout an enterprise.
• Workload Independence: A workload acceleration engine should be able to improve the performance of any workload regardless of the application or amount of data being moved over distance.
• Zero Touch Installation: No need to configure complex networking in-line protocols on the physical equipment, such as routers, switches, and firewalls. Acceleration is enabled right from the management console. (See Figure 3)

Use Cases for Workload Acceleration

Workload acceleration benefits any distributed organization that has virtualized their infrastructure with the following being some common use cases:

Application owners and virtualization managers within large enterprises must work with the networking team to accelerate their traffic over distance. This can involve meetings and bureaucratic delays, not to mention time to deploy and test equipment. Workload acceleration eliminates much of that delay by letting virtualization managers deploy acceleration with point and click simplicity.

Companies with managed WAN services and cloud deployments often want to accelerate an application but are unable to access the network equipment owned by the service provider. As such, they cannot make whatever modifications may be required to deploy an acceleration engine. Workload acceleration improves the application’s performance without touching the service provider’s equipment.

A New Level of Agility

In an agile business, application owners and virtualization managers need to deliver solutions fast. This requires new functionality within the virtual environment for accelerating workloads across a distributed environment. Workload acceleration gives virtual administrators the performance, flexibility, and control they need. By enabling these individuals to maximize the performance of individual workloads in a fast, easy and highly-efficient manner, workload acceleration is a key requirement for any virtual enterprise.