SCO-Overview

I have to admit, I’m a bit excited to get my hands dirty with IBM’s latest release of SmartCloud Orchestrator (version 2.2) which was just released this quarter.  With the new tools they have developed we have now a series of features that can actually deliver on the promise of easy to use orchestration. But before I dive into the details, a little history from the F5 perspective.

Step into the Wayback Machine™

Here at F5 we have always been agnostic about our orchestration partners. Our goal is to have a series of open, stable and well documented APIs that anyone can use to program our devices and software modules. To this end, iControl has been the API of choice for a decade.  iControl has been a SOAP based interface.  In version 10 we also added the TMSH, the Traffic Management Shell for command and control from the command line interface (CLI). TMSH is accessed through SSH and has full contextual help and autocomplete.  Now, with BIG-IP version 11.4 (released June 2013) we have also introduced iControl-REST. This now allows the remote API based control of BIG-IP through the web services model of “Representational state transfer” or REST.  Exciting stuff, for people that care about this sort of thing (and everyone should care, really) because automation is a beautiful thing that can reduce operational costs and increase uptime and visibility.

So, there’s the quick history lesson.  In regards to IBM, our great partner has adopted multiple models.  There is the ability to use the SOAP interface, of course, and IBM has also adopted TMSH through two different offerings, first, the Tivoli Service Automation Manager (TSAM) and second, the Netcool Configuration manager.

You see where this is all going right? F5 has built a base of open, stable and documented APIs, there are a variety of tools customers can use to orchestrate the BIG-IP, but so far it’s been rather complicated to decide how to best do this. A number of orchestration tools have come and gone from a variety of vendors but this brings me back to the beginning, I’m excited because I think IBM has got it right by packaging together all of their expertise and knowledge into the latest set of tools.

What are the new tools?

IBM’s focus on SmartCloud offerings has led to the creation of an open and scalable cloud platform, with easy to use orchestration and service automation and a marketplace for automation packages.  And these products at the core are SmartCloud Orchestrator and SmartCloud Provisioning.  I believe that the orchestration features, as they relate to F5 and BIG-IP, go beyond just cloud use cases and apply to the enterprise as well as the private, hybrid or public cloud.  The provisioning features make it easy to deploy these in a true cloud infrastructure. Let’s look at a couple of examples, one in the enterprise and the other in a SmartCloud setting.

In the enterprise scenario, SmartCloud Orchestrator and the Content pack would be deployed internally. From there, it’s all about creating the connection to BIG-IP and building the routines that need to be executed. This may be as simple as adding new pool members as demand warrants, or as complex as you can imagine.

You can find documentation on the content pack for the F5 ADC here: ftp://ftp.software.ibm.com/software/tivoli/OPAL/1TW10SO06/IBM_SCO_Content_pack_F5_Load_Balancer.pdf

You can find the actual content pack download here:

https://www-304.ibm.com/software/brandcatalog/ismlibrary/details?source=cloudmarket&catalog.label=1TW10SO06

And you can try a hosted beta of Orchestrator here: https://www.ibm.com/developerworks/community/wikis/home?lang=en#!/wiki/W14a769bd540e_4b7c_aefa_cf851f2221a7/page/Using%20the%20SmartCloud%20Orchestrator%20hosted%20beta%20system

In the private cloud, Smart Cloud Provisioning and the Content Marketplace would be used to deploy a tool such as Netcool Configuration Manager for SmartCloud Enterprise and then you’d be off to the races, creating the automation tools.

So, IBM has brought together all the APIs and tools under one umbrella that promises to be easy to deploy internally or in the cloud.  But does it work? And how easy is it, really? I’m setting it all up in my lab right now and I’ll be reporting back in a matter of weeks with my findings.  Stay tuned….