Networking is no longer restricted to the hardware gear of routers and switches as organizations latch on to software-defined networking. Juniper Networks catering to large enterprise, enterprises and service providers too, is changing its GTM in the cloud-centric, software-defined world. IDG India spoke to Dinesh Verma, Managing Director, India & SAARC, Juniper Networks on the company’s India outlook, emergence of SDN and CIO priorities.
What’s the maturity curve of Software-defined Networking (SDN) amongst Indian organizations? Industry reports suggest that it’s still in the hype cycle.
The key question is to make use of this technology in right manner. I agree SDN is bit of hype and a bit of practicality and hence we are concentrating on the practical use of SDN technology for Indian organizations. The early adopters of SDN have definite benefits. These are the organizations with thousands of locations that need to be connected and secured.
However the major problem they face is the issue of security on their SDN journey. Whether it should be domestic or international? Whether the bigger control in the entire topology is robust? Can IT team manage the whole infra like it past two decades? These are few questions occupying minds of CIOs’ who are contemplating to adopt SDN. The ownership lies with tech vendors like us to convince them of providing the robust architecture, advocating the use case and designing the solution as per that vertical and its business need. Many companies are failing left, right and center with SDN because I think that the big gap actually lies between the use case and the adaption. Therefore we are threading SDN carefully in India and not just pushing it as a new technology.
Any particular verticals that are latching onto SDN faster than others?
In fact there are conversations across the verticals about SDN. Whether the hardware-software desegregation will happen and when, are open questions being talked in the industry. The companies will start small, start solid with SDN and that's the way to succeed. Otherwise you will kill the technology the way 3D, for example, got killed in India.
One needs to be very careful advocating the use case of technology. Any insurance company, BSFI or large retail company are all the right candidates for software-defined technology or companies with global operations. In far-flung areas, the factories or the organizations with distributed operation are the right focus for SDN. The larger the number of offices, they become better potential. It depends on whether the technology is proven for that particular industry and the confidence of the vendor in its tech portfolio than making a quick buck behind SDN jargon in the market.
Juniper’s SDSN (Software-Defined Secure Network) as a cyber security platform is an integral part of the SDN story. Our Virtual Chassis port (VCP) with VSRX become a pretty solid proposition for companies thinking of software-defined journey. Juniper’s vSRX Virtual Firewall delivers a complete virtual firewall solution, including advanced security, robust networking, and automated virtual machine life.
What do CIOs and CISOs at the customer end expect from network companies like Juniper today versus three years back?
It’s a very interesting transition for IT decisions makers. If I had to be impartial to them they are still in the adoption curve which has not taken off for most of them. All CIOs/ ITDMs and all tech vendors are talking about cloud, software-defined, virtualized platform but I am not sure if everybody has actually understood the nitty-gritty of the entire ecosystem. That's the problem statement with most of CIOs as they are not very sure to jump on SDN bandwagon. They are uncertain of the seamless migration to the new platform, and smooth operations on the new platform; but they all believe that there will be more opex and reduced capex in terms of IT investments. TCO from SDN will happen over the next five years; but somebody at the customer end needs to take the lead to make the transition to SDx. They want someone (vendor) to ensure the migrations from A to B platforms and someone to hold their hand for say next six months to 12 months.
If you look at TCO over 5 years for SDN solutions, the benefits are there on the wall as the opex will cut down by 40% to 60% and it's not difficult to prove the savings to a CFO. The CFO will approve the project; but usually the bottleneck lies with IT leader’s and his team’s confidence to make it happen. That hurdle has to be crossed which we are trying to break through our solutions’ value prop and hand-holding them on the journey.
That means there is too much of pre-sales by Juniper team required for SDN to move faster.
That’s right. Very heavy pre-sales and professional services because it has to be done right at the first time. We thoroughly give them the blueprint on what they experience today and near future which is at a scale and that is manageable. The early adopters just went ahead with the flow and some of them succeeded though it was risky being too ahead of the curve. The concept on paper versus on the field for SDN are totally different because otherwise, Juniper or any of our competitors wouldn't have spent decades in perfecting the hardware platforms or the software platforms. Even the hardware-software desegregation takes time though you can always announce it much in advance.
What about the traditional hardware business that is the ‘bread and butter’ for networking vendors?
That business is growing not only across enterprise segment in major verticals but our key focus segments like cloud or service providers. We are still one of the leaders in the business of main routing and core routing. The adoption of hardware switches is very strong in India with a good growth trajectory.
The hardware appliances typically are not brought by the companies for their own data centers but the networking gear is now present in the third-party datacenters. That’s the shift for many years now. We are developing data centers for many of our very large enterprise customers also and those are very large data centers through Spine-and-leaf technology, EX LAN Series as examples. Juniper VPN which are open standards are gaining traction more than the closed architectures that many of our worthy competitors propagate in the market.
What about Juniper’s collaboration story with other industry vendors?
The new announcements couple of weeks ago has all bits and pieces all the way across as we have merged together along with our partners which are primary and open source to create a solution offering which allows the enterprises or the cloud operators to run a secure multi-cloud operation.
The security of multi-cloud operations is inbuilt at our end. Multi-cloud means whether private cloud or using public clouds like Amazon, Google, and anybody. Or they have their own data centers spread across the world to do load migration from point A to point B which is transparent to the user.
Hence it’s not required to know where the load is working or whether the application is working as companies can take control, and it could be either on a bare metal or a virtualized platform so that particular offering that has been stitched together now is an extremely powerful tool and that works on open standards. We work very closely with Red Hat, Veritas and we have our own software that enables this ecosystem. It’s a very powerful story not for the enterprises but for the telecom cloud where multi-cloud architecture within the enterprise itself.
But multi-cloud strategy can create complexity in the company’s IT infra or cloud roadmap.
Why should there be complexity? When we talk about multi-layer straight away we go to Amazon and other platforms? An enterprise has ten cloud data centers of their own with different architecture through different OEMs. So how would you control that? How will it create a multi-cloud architecture to use these data centers to any of the cloud infrastructures and control it as one platform where the loads shift from one Software-defined datacenter to another? Companies prefer to go to Amazon or anything because you get all the apps readymade and you can use it as you grow. Multi-cloud doesn't necessarily mean that you go into public cloud but it could be your own. Like a large bank with multiple data centers doesn’t need to create a primary and a backup for their core banking application, but they can distribute it across to the nearest branch with least latency to that particular exercise.
Who are the actual buyer of your technologynetwork engineers, CIOs or LOB’s or emerging CDOs? Highlight your India priorities for 2018 ?
We have not seen much demand from CDOs but definitely it is more from CSOs and CIOs. The CIO and CTO depending on the organization are the two major positions of influencers and in the case of multi-cloud and SD-WAN, the LOBs also play an important role in tech buying.
When I came on board our dependence on telco was the highest but today we have branched out into various enterprise segments and also government. This has made our business more predictable. Nobody does security the way we do, nobody does multi-cloud the way we do and whether it is underlay or overlay, the new topologies and the new upcoming architectures which will be more useful for Telco or BFSI in India. We don't see much of manufacturing; but telco and BFSI along with IT/ITes where they are migrating workloads from on site to cloud hold good potential for our new portfolio.
The business climate is healthy with 2019 being the election year in India, across enterprise and government segments investing in technology. Organizations taking the leap ahead to catch wave of disruption in terms of adoption of new tech will stay ahead. This will happen now more across companies because the maturity curve has progressed a lot versus two years ago. We will put all might to evangelize the technology which will be more multi cloud-based, more software-based and more security-based. Data security is any which ways very important, and mobile 5G expected to take off in 2020 will require edge computing technology. We will be well placed for the edge computing too.