F5 Friday: How Can I Manage Thee? Let Me Count the Ways…

#v11 A robust and diverse set of management tools enabling a variety of infrastructure integration options is essential to architecting a dynamic data center

In the continuing quest for a more dynamic data center, infrastructure integration must necessarily take center stage. While virtualization has enabled fluidity of server infrastructure, it has not done so for the network and may never be wholly suitable for the task for a variety of reasons.

But the agility resulting from virtualization, the ability to manage resources on-demand, must be incorporated into the network infrastructure in order to scale an on-demand, dynamic data center. Without the means by which infrastructure can adapt to the changing allocation of resources it remains a static beast requiring additional time and effort to manage in real-time.

It is not a scalable model. The network must adapt, it must become more fluid and capable of reacting to events and triggers and conditions with respect to the security, performance, and resiliency of the applications it is ultimately tasked with delivering.

But because no two data centers are the same and because the introduction of cloud computing has, well, made architectures even more cloudy, it is necessary for infrastructure to support a robust management plane, capable of integration with other data center components, management systems, and third-party frameworks. F5 has long been committed to a robust and flexible management plane, and the introduction of iApp in BIG-IP version 11 only expands on existing management-focused technologies.

THE F5 MANAGEMENT PLANE

The F5 BIG-IP management plane comprises three categories of technology through which BIG-IP and can be configured, controlled, and integrated.

LEGACY

A stable command line interface (CLI) is the staple of network infrastructure component management and F5 BIG-IP is no exception. F5’s CLI allows complete command and control of BIG-IP technologies across the core system and its diverse set of application delivery modules. Also falling into the legacy category is SNMP support. F5 maintains a set of SNMP MIBs enabling remote management and monitoring from virtually any SNMP-compliant management system or framework. SNMP support is key to supporting F5 integration with many network monitoring and management frameworks such as Cacti  and Nagios .

API BASED

F5’s API-based management is a programmatic means of control and integration. Introduced in 2001, iControl provides fine-grained control over nearly every aspect of both execution and configuration on BIG-IP systems. With specific APIs designed for use with each module, iControl is a robust, flexible means of providing control over F5 BIG-IP. It’s SOAP-based interface makes it interoperable and the DevCentral team currently supports a wide variety of libraries and assemblies specifically aimed at making use of iControl as easy as possible. Python, Microsoft PowerShell, Java, Ruby, and even PHP are among the languages and environments supported.

iControl is the primary means through which F5 BIG-IP is integrated into both commercial management offerings and cloud computing environments. Oracle, Microsoft, WhiteHat Security, HP, and VMware all leverage iControl as a means to integrate BIG-IP and its application delivery services into their respective offerings. Cloud computing providers such as Rackspace leverage the API to integrate BIG-IP scalability services into its offering, making load balancing services easily accessible to its customers.

Not to be overlooked is F5’s own Enterprise Manager: a stand-alone management appliance (also available as a virtual edition) that, like our own partners, leverages iControl to manage BIG-IP systems. It is designed to provide basic configuration management of multiple BIG-IP systems and uses a centralized model to provide a holistic view of performance and configuration of BIG-IP devices across the data center.

More on iControl:

SCRIPTING BASED

As the role of devops within organizations continue to develop and evolve into an operational must-have, scripting-based integration of network infrastructure becomes vital to the design of repeatable, successful application deployments. In the past, TMSH (TMOS Shell) has been the primary remote (and local) scripting language of choice. TMSH offers an object-based scripting paradigm leveraging a TCL-like interface providing much the same level of configuration and management as iControl. TMSH requires less development-focused skills and knowledge to leverage, and integrates easily within the current script-based devops paradigm for configuration and management.

With the introduction of BIG-IP version 11, iApp has joined the management plane as an evolutionary scripting-based option. The core difference between TMSH and iApp is that iApp is more abstracted and includes a user-configurable, GUI template system that is more in line with current thought on the development of self-service options for infrastructure. iApp enables packaging of application delivery service configuration in a way that provides for repeatable, consistent deployment of applications architectures – not just applications. This capability aids in reducing time to deploy and the rate of human error introduced into lengthy, complex manual processes. Managing BIG-IP application delivery services via iApp enables operations to define services that can be self-service configured by IT or even business users as well as integrated into operational process workflows through scripting or other event-based triggers. iApp is a completely new way to manage infrastructure, and it moves the management of application delivery closer to the paradigm of self-service, automated networking frameworks similar to those leveraged today in cloud computing and highly-virtualized environments.

ONE SIZE DOES NOT FIT ALL

The premise is that no single management solution can support every customer with equal satisfaction. There will always be exceptions to basic management system rules, and there will always be custom integration requirements – more so now that enterprise IT is looking to enable a dynamic data center and ultimately IT-as-a-Service. Whether it’s integrating with Puppet or Chef in a devops-focused model, pre-integrated with VMware or Oracle, or a custom integration effort to enable infrastructure as a service in the data center or as a cloud computing provider, F5’s management plane provides the flexibility necessary to manage infrastructure, as a service.

The need for integration at the infrastructure layer is not going away and, in fact, will escalate as organizations turn their eye-of-efficiency toward the network. The ability to automate network components such as application delivery controllers and enable self-service management for specific application delivery services such as load balancing, acceleration, and security will become paramount to a successful drive toward IT-as-a-Service. A robust and flexible management plane is a must-have for infrastructure components to support and enable the integration necessary to automate and orchestrate data center operational processes.

Happy Managing!

 

AddThis Feed Button Bookmark and Share

Published Sep 16, 2011
Version 1.0

Was this article helpful?

No CommentsBe the first to comment