I recently had a conversation with Prabhu Ramachandran, senior director
of WebNMS, a division of Zoho Corporation,
where we talked about how service providers are improving their networks with orchestration.
VMblog: The entire
industry is talking about orchestration. Why is orchestration important to
service providers? And what is WebNMS offering in this area?
Prabhu Ramachandran: The end goal
of orchestration is service agility - delivering services more quickly and
using automation to control operational costs. At WebNMS, we released the Symphony
Orchestration Platform to address this need. Symphony is a suite of
software solutions and tools that give service providers a foundation to build service
orchestration applications that automate their operations. It helps providers
deliver much more dynamic services to their customers and reduce their OPEX.
VMblog: How is Symphony
different from other orchestration products? What do you offer that's unique here?
Ramachandran: From an
operator's perspective, Symphony is unique in that it combines true workflow
customization with an unmatched library of multi-vendor network management
solutions in an open environment. Symphony focuses on helping providers
automate existing, well-defined services to get OPEX efficiency now and prepare
for SDN and NFV, with open APIs that prevent vendor lock-in.
Symphony leverages the WebNMS Framework, a field-proven,
carrier-class network management solution, with over 400 service provider and
system vendor customers. This eliminates many integration headaches, unifying
their network management, which we see as a first step on the path to orchestration.
We want users to focus their efforts on workflow automation.
Third, to make those workflow automation efforts more tractable
and efficient, we've applied our experience with software frameworks to deliver
a purpose-built tool for custom workflow automation. Providers can efficiently
create complex workflows, either themselves or working with a system
VMblog: How do you define
Ramachandran: First of all,
we chose the word ‘workflow' because people can relate it to their day-to-day
activities, which we are trying to make more efficient through automation.
To us, a workflow is a representation of a process in
sufficient detail and completeness that it can be executed by automated
Workflows include the inputs and outputs of the process as
well as the steps of the process itself. One important input is the state of
the network, which is often very complex and brings about the need for big data
tools as part of workflows. Workflows often modify that state, say by
configuring a new service. They also deliver outputs like reports,
acknowledgements or a list of issues that need resolution.
Symphony helps structure workflows, which improves
productivity, but creating workflows still takes effort.
First of all, workflows need to be complete for a scope of
work. In our manual processes, after we do them a few times, we can learn to
take completeness for granted. Think about teaching kids something new. How
many times do you leave out a step? Correct workflows capture all the steps, no
matter how obvious.
They also must be repeatable - or automated. You get ROI
from the Nth iteration, not the first. When the workflow is invoked by some
event, it should produce the desired results the vast majority of the time. One
hundred percent success would be great, but sometimes the incremental work to
describe that level or completeness is not worth it, at least not right away.
Practical workflows often need exception paths to handle rare cases - perhaps
escalating an irresolvable performance issue to a NOC engineer.
I like thinking about printers as a good example of workflow
optimization. Printers are part of our workflows. Having a home or office
printer is infinitely better than having a typewriter or getting film
developed, but I still run out of paper or ink sometimes. Adding paper every
few days is an acceptable trade-off for me. Of course, the printer companies
are trying to improve their customer experience through improved automation - they
warn me when my ink cartridges get low and will place an online order for new
ink. But, for me, that's a diminishing ROI compared to the ease of printing
directly from applications.
We are very hopeful about many workflow possibilities for
networks (i.e., ordering new infrastructure, particularly NFV resources based
on load), but we feel the biggest near-term return will be in service
provisioning and assurance. Providers need to make that dramatically more
efficient because it happens all the time, and end users want to get new
VMblog: Are your customers
automating service assurance workflows?
Ramachandran: Yes, service
assurance is a common thread. It's critical to provider customers, and it can
be a time-consuming task, particularly when you have a large installed base of
For example, we've partnered with VeEX to automate remote
service activation testing, initially for Carrier Ethernet. When establishing a
new service, a workflow provisions the service and then instructs the remote
test equipment to verify the expected performance with a live network test. The
benefits here are you get an assured service without a truck roll - saving time
Then, once services are established, another workflow can
monitor all of the services continuously and ensure they are performing within
acceptable limits. If performance drops off for one or more services, the
workflow can raise an alarm for resolution.
VMblog: You also mentioned
big data. How about big data use cases?
Ramachandran: Right now, big
data's primary benefit is reliability through performance at scale.
Orchestration centralizes information processing. All the
workflows share a common database of real-time information to stay in sync.
This database grows over time and can get "big" pretty quickly. Raw size is one
stress on performance, so the scalability of big data is helpful.
But, performance reliability when the network itself is
under stress is even more important. When something goes wrong in the network,
it can generate a flood of alarms. Workflows issuing and processing of these
alarms generate spikes of database operations. It can overwhelm conventional
databases. If the database performance lags, the operator can't trust the
underlying automation and must fall back to manual operations at the worst
possible time. Our big data repositories are showing dramatically improved
performance - removing the database as the bottleneck and providing headroom
for business growth.
Over time, we expect providers will find many creative and
proactive uses for big data analytics. An analysis of the load on a business
user's service might show a service usage saturation trend and allow the
provider to inform the customer that a service upgrade is needed to avoid
performance problems. Particularly for fast-growing SMBs, likely without their
own sophisticated IP, getting that warning through a customer portal could
avoid slowing down their business - adding value to the provider's service.
Our customers see the potential for innovation
here. Today, most are focused on streamlining their current operations but want
a platform that can integrate these capabilities once baseline automation is in
With over 14 years of experience delivering service
provider software solutions, Prabhu Ramachandran directs WebNMS, the service
provider division of Zoho Corporation. Prabhu leads strategic marketing,
product management, customer support, partnerships and professional services
for WebNMS. Leveraging the technology of the corporation's flagship WebNMS
Framework, Prabhu has expanded the business from its longstanding leadership
position in multi-vendor network and element management software into vertical
solutions for Carrier Ethernet, MPLS, broadband, LTE and satellite networks. In
2012, Prabhu began driving WebNMS into network orchestration, SDN, NFV and IoT/M2M
platforms, all critical enablers for service providers to grow profitable
businesses. He holds a Bachelor's Degree in Electronics and Communication from
Madras University, Chennai, India.
For more information about WebNMS, please visit http://www.webnms.com, and follow
the company on Twitter at @WebNMSIoT.