Cloud migration brings automation and programmability to the forefront

By Greg Barnes, A10 Networks ANZ managing director

When organisations move their application infrastructure to the cloud, they need to rethink everything they ever learned about application networking.

Traditional network architectures based on static, single-purpose appliances will prevent them from realising the benefits of cloud computing.

Cloud, and initiatives such as continuous application deployment, hyperscale networking, mobility and big data, all require automation. In turn, automation requires programmability.

And ultimately, programmability using RESTful APIs allows administrators to quickly integrate application delivery platforms with other services and management systems, thus avoiding costly and protracted integration projects.

Additionally, as cloud networks mature they grow richer and more heterogeneous. They incorporate all the services and management tools of traditional network architectures, and this ecosystem of solutions helps to improve the uptime, visibility and security of cloud deployments.

To ensure that this ecosystem of services works together harmoniously, application networking platforms need to support open and standards-based programmability.

All aspects of management and monitoring should be accessible from standards-based APIs – providing interoperability with automation, orchestration and analytics tools.

DevOps and continuous deployment

In rapidly evolving cloud environments, developers and IT teams cannot be siloed; they must collaborate to accelerate and improve the quality of product releases.

DevOps practices aim to improve communication and trust between software and IT departments either by creating a new cross-functional team within an organisation or by facilitating collaboration between developers and IT operations to increase productivity.

The DevOps movement is characterised by frequent application releases that allow organisations to innovate swiftly and adapt to customer and market demands.

It also incorporates techniques that reduce errors and shorten the mean time to recovery (MTTR) in the event of an application failure. Incorporating elements of agile software development, DevOps improves version control, tracking and testing of releases to reduce errors.

But DevOps depends heavily on automation to achieve its objectives. DevOps teams cannot hope to deploy new application builds multiple times a day or spin up and spin down servers manually.

Engineers must be able to provision software with the click of a button. Virtual switches and load balancers in the cloud must dynamically adapt to these application changes.

If new applications are deployed, then application delivery software must distribute traffic to them without requiring time-consuming and error-prone manual intervention. Ideally:

• DevOps teams must have visibility into application performance.
• Monitoring tools must alert on application problems and security threats.
• Automated reports should provide a centralised view of network operations over time, revealing trends and helping organisations plan for future growth.
• Application delivery solutions should support automated provisioning and integrate with cloud orchestration systems.

SDN requirements

Like other cloud computing trends, SDN requires programmability. Since SDN decouples the data plane and the control plane, the data plane must be directly programmable and it must support open, standards-based APIs. Network operators must be able to configure, manage and monitor network resources centrally.

For rapid SDN and network functions virtualisation (NFV) integration, networking platforms like load balancers and switches should support open and vendor-neutral APIs.

By supporting standards-based programmability, application delivery controllers (ADCs) can allow DevOps engineers to write their own policies or provisioning scripts themselves. This empowers DevOps engineers to tailor automation policies for their application needs.

Automating security

As organisations migrate their application infrastructure to the cloud, security defences must migrate too. Unfortunately, cloud environments can introduce complexities that often force DevOps and IT security teams to re-evaluate their security strategies.

To prevent the onslaught of today's security threats, DevOps engineers must work with IT security administrators to deploy DDoS protection as well as Web Application Firewall (WAF) and DNS Application Firewall (DAF) technology to secure cloud infrastructure.

In cloud environments, security solutions must be able to adapt dynamically to network changes. When an application workload is added, security defences must be automatically enforced, ensuring that the application is protected against attacks.

Whether deploying applications in virtual private clouds or public clouds, the security and high availability solution must be able to scale out on demand and integrate with existing management systems. Only then can organisations enforce consistent security policies across all data center assets.

Next generation-ready architecture

Emerging architectures represent a new standard for programmability, policy control and telemetry that satisfy the application networking requirements of the cloud and provide a rock-solid foundation for the future.

Select an architecture that supports instant integration with SDN frameworks, toolsets for policy enforcement, and comprehensive counters and APIs for visibility. The architecture must also allow organisations to consolidate services on a powerful and efficient application networking platform.

A new-generation architecture will allow developers to easily integrate intelligent application networking with selected application delivery controllers, laying the foundation for software-defined application delivery. With open, programmable interfaces, DevOps administrators can integrate services faster and apply consistent policies across the data center.

If organisations choose to use their own management system or want to customise an existing orchestration tool, they do not need to invest time and money in integration efforts or professional services projects. The right application delivery controller will offer full access to all configuration settings through standards-based RESTful APIs.

Organisations can leverage an appropriate ADC's standards-compliant programmability to integrate and provision services and automate management processes even faster. In today's cloud environments, integration with SDN controllers and orchestration systems is not optional – it is a must have.