Software-defined data-center (SDDC) networks hold the promise of quickly and automatically reallocating resources to best support applications without changing the underlying physical infrastructure, but they require the proper integration of management, automation and network orchestration (MANO).
The SDDCN combines with compute resources (virtual machines and containers) and storage (disc and flash) to deliver specified performance for private cloud applications
MANO represents critical elements for making private clouds operate smoothly in a multi-cloud world. Each MANO element in the SDDCN has unique characteristics and challenges as outlined below. These definitions overlap and are not mutually exclusive.
SDDCN automation
Network automation is defined as the ability to replace previously manual processes via software tools. Software can be deployed in the data center to automatically configure and provision network resources – for example, vLANs for new container-based applications. Automation is critical for the SDDCN to handle the vast number of micro-services and traditional applications in most large data centers – each with specific network and security requirements. Automation benefits in the SDDCN include the rapid provisioning of network services and the reduction in network issues created by human error. Hyper-scale cloud providers like Google employ network software to almost completely automate provisioning of network services in their enormous data centers.
Data-center network management
Network management is a critical process for administering, operating and maintaining high-quality network flows in the data center. Network operators are responsible for delivering uninterrupted quality of service for all applications resident in the data center. When the network is down, compromised or unable to deliver expected quality of service, network administrators need processes to troubleshoot the problems and to rapidly remediate any issues. Network-management software is responsible for keeping an accurate inventory of all devices connected to the network and for tracking traffic flows within the data center and to external sites such as public clouds, SaaS and remote data centers.
Orchestration
Orchestration provides a policy-driven approach to network automation and management. It can program the network to deliver the appropriate compute and storage resources to each application. Leveraging the principals of SDN, network orchestration automates the method by which applications request network resources and helps minimize human intervention. Some examples of network orchestration include using traffic engineering to provide appropriate quality of service and providing security policies by user and application. Increasing network orchestration will be required to enable organizations to leverage multi-cloud architectures and run applications across a variety of architectures on-premises and off-premises.
Intent-based networking vs. SDDCN MANO
Intent-based networking is the new buzzword that incorporates many elements of SDDCN MANO as described above. It abstracts network complexity and improves automation by reducing or eliminating the need for human intervention. The goal is to enable IT personnel to use natural-language requests for allocating network resources. For example, IT could request via a simple software command increased performance for the CEO’s video conference. Current SDDCN technologies can automatically assign IP addresses, configure vLANs and provide insights into quality-of-service challenges. The goal of intent-based networks is to leverage artificial-intelligence capabilities to smoothly run a complex SDDCN with limited need for manual processes.
MANO challenges in data centers
Many suppliers are marketing newly added data-center MANO or intent-based capabilities. These solutions do allow data-center network administrators to largely eliminate the need for traditional methods of network management via command line interfaces, which is cumbersome, slow, labor-intensive and prone to error.
However, most suppliers of intent-based network software are still in the early phases of their innovation. There are only limited standards for SDDCN MANO and associated SDN controllers. Most network administers will continue to rely on mainly single-vendor solutions with tight integration between data-center-network software and hardware.
SDDCN MANO examples
Cisco and VMware both separately announced ambitious plans to enhance their data-center networking solutions via improved automation. A number of smaller suppliers are innovating around intent and network automation in the data center. Some examples include:
- The Apstra Operating System provides control and visibility for multi-vendor network elements in enterprise data centers.
- Big Switch’s Big Cloud Fabric provides network integration and visibility for multi-vendor environments including OpenStack, containers and hyperconverged systems.
- Cumulus Network OS enables customers to automate deployment of a range of white-box switches and to leverage Linux-based tools to customize operations and management.
- Forward Networks analyzes data-center network behavior to enhance analytics and provide actionable insights.
- Gluware offers a platform for network configuration automation and monitoring for a range of multi-vendor network elements in the data center.
(Note: Cisco, Cumulus, Gluware and VMware are clients of Doyle Research.)
Recommendations for IT executives
The scale and complexity of modern data centers require advanced network software to provision, manage and secure high-speed traffic flows to VMs, containers and legacy applications. Network administrators need automated solutions to monitor and deliver reliable quality of service to critical applications. Thus, MANO is a critical element for selecting appropriate network partners for the data center.
IT organizations still face the challenge of whether to base their data-center network on one vendor, select established suppliers based on best-in-breed capabilities or work with innovative start-ups. IT leaders should evaluate SDDCN MANO solutions based on the following:
- Network-automation capabilities
- Ability to monitor traffic and enforce QoS priorities for critical applications
- Network intelligence to identify and block anomalous traffic and route suspicious traffic to dedicated data-center security systems
- Ability to migrate and track application/workflows in multi-cloud environments, on premises and cloud-based
- Support for multi-vendor environments and depth of network/data-center-partner ecosystem
Suppliers can be judged on their current capabilities and their ability to deliver MANO improvements on a one- to two-year road map.