The past year was a busy one for the networking industry, with lots of promises for great things to come starting in 2014.
It should be another busy and dynamic year. Here are the important trends we think are worth watching for.
Use Cases: Network Virtualization, Bandwidth Calendaring, and Network Tapping Lead Early Deployments
While that killer use case for makes software-defined networking (SDN) and network functions virtualization (NFV) remains as elusive as the Alaskan blue crab, three important use cases are making progress into customer networks.
Two groups of customers and use cases are driving network virtualization: cloud service providers and enterprise data centers.
Most cloud service providers are somewhere between early trials and limited production with their network virtualization deployments. The need to offer either competitive data-center services (i.e., compete with Amazon) or build large-scale computing for their own use (Facebook) will make network virtualization an important topic in 2014. What to watch:OpenStack and non-VMware providers.
Enterprise customers we speak with are looking to use network virtualization to cost-optimize their legacy networks and data centers. Typically, these are regulated enterprises, predominately using VMware, that are looking to consolidate compute and storage resources that today are separated by firewall. VMware’s NSX appears to be a logical (and easy) choice to serve this segment of the market. What to watch: NSX traction; other vendors with network virtualization compatible with VMware, like Plumgrid.
Bandwidth Calendaring (for WAN-link optimization)
Two years ago, Google talked about how they customized OpenFlow to optimize the WAN via bandwidth calendaring. Today, every major telco, MSO, cloud service provider, and data-center provider is looking to optimize transport. Look at customers such as AT&T and Microsoft to lead the requirements and for vendors such as Cisco, Ericsson, Ciena, Cyan, Alcatel, and Juniper to lead figuring out how deployment this in world-class multi-vendor networks.
Today, some of the most advanced customers we speak with tap 8 to 10 percent of their network ports, and they articulate the value of being able to tap all of their network ports. They also tell us the current architecture of building parallel monitoring networking to implement tapping is too expensive. This is driving interest in leveraging SDN to deliver networking tapping over a virtual monitoring network that runs on top of a production data center fabric. What to look for: Check out traditional monitoring vendors such Gigamon and VSS Monitoring as they expand their SDN offerings; look at new capabilities from Cisco and Arista (among others), and start-ups like Big Switch Networks.
Year of the PoC
2012 was the year for people to learn “What is SDN?” and 2013 was the year for people to understand how they use SDN.
2014 will be the “show me” year, the year of the proof-of-concept (Poc) for SDN, NFV, and network virtualization. We’re still at least a year away from significant production deployments, but we expect to see a large number of customers installing PoCs in their labs. Watch for PoCs from non-technology organizations such as major service providers and financial services firms.
Security: Top Driver and Inhibitor for Data Center Network Virtualization
Customers are telling us they need to prove that a virtual network is just as secure as a physical network. At the same time, the key driver for enterprises to move to network virtualization is to reduce the number of physical security appliances. What to watch for: startups and incumbents like Palo Alto Networks, Cisco, Netskope, and Illumio duking it out to solve this problem.
DevOps: New Key Technical and Economic Buyer for Networking
The hottest title in IT today, DevOps is breaking down political barriers and creating new power centers within the IT shops. Watch how buying patterns and purchasing change within IT organizations. Network architects and engineers: Look to expand your knowledge of non-networking technologies so you are not left behind.
Open-Source and Open Networking: Beware of Proprietary Software in Open-Source’s Clothing
“Open” has turned into a major area of debate and discussion across the networking community with every interested party (customers, suppliers, researchers) all proposing their own definitions of “open.” While much of the current debate has centered on open vs. closed-source approaches, we expect this debate to shift in 2014 to which approach(es) to “open” best solve customers’ technical problems and business requirements while enabling suppliers to build sustainable businesses.
Watch to the end of 2014 for general consensus around what open networking is and the role of open-source. We’ll start to see a number of different ecosystems emerge around different open networking approaches.
White Box Networking: Two Camps, Too Early to Call a Winner
Two camps emerged in 2013 around white-box networking. One believes that large service providers, data center operators, and other organizations whose existences depend on networking technologies are demanding white-box networking gear, similar to evolution of white-box servers three to five years ago. To subscribe to this thinking, one needs to believe the customers fundamentally want to replace their current suppliers and that open-source networking technology is good enough to replace proprietary systems.
The second camp believes that white-box networking may limited to the biggest network operators (Google, Facebook, Amazon, Microsoft, Yahoo, AT&T, Verizon, etc.). The feeling here is that only these massive operators have both a) sufficient software development and network operation capabilities and b) scale to justify the financial investment in white-box networking. The view here is that everyone else may use white-box networking to extract better pricing from their current suppliers.
Watch to see who adopts white-box networking beyond lab trials and backwater use cases.