Credit: Shutterstock photoI attended and participated on various panels at the Carrier Network Virtualization conference in Palo Alto from December 9-11, 2014. This conference was the fourth in a series of Software-Defined Networking (SDN) and Network Function Virtualization (NFV) conferences that began in Barcelona in June 2013. I encourage readers to review my previous SDN and NFV conference articles from June 2014 , December 2013 , June 2013 , and SDN literature from April 2012 as they provide an evolving perspective on this new telecom infrastructure paradigm.
As a primer, SDN is direct programming that tells the silicon switches what to do. An SDN network can be managed as one virtualized switch, instead of trying to manage many switches and routers all over the network. This is accomplished by a separate control plane that can route network traffic ideally with one standard protocol (e.g. OpenFlow) as opposed to thousands of hardware routers with their proprietary network protocols. Below the control plane would be a forwarding (a.k.a. data) plane consisting of high-speed, lower-cost switches, as the intelligence rests at the centralized control plane.
NFV is similar to SDN but with more of a telecom focus that is less tied to the OpenFlow protocol. Similar to SDN, NFV focuses on virtualizing the separate network hardware appliances like routers, firewalls, load balancers, and distribution switches that are all over the network into a centralized virtual switch that incorporates all these hardware functions. An Orchestration layer is then needed to create a total solution that can interoperate with open-source and carrier-specific protocols.
From a vendor perspective, the bulk of the SDN and network virtualization market today is in the data centers. According to Infonetics, the software-defined data center (SDDC) is a $1 billion market today, almost totally consisting of Web 2.0 data centers like Amazon ( AMZN ) and Google ( GOOG ). The SDDC market is projected to jump to $15B by 2018, two-thirds from Web 2.0 data centers and one-third from carrier data centers. All of these figures represent a sliver of the $354B global carrier capex estimated for 2014. The non-data center, carrier SDN/NFV market is nearly non-existent today, but has the most potential over the long run driven by network virtualization software.
I organized and chaired a panel featuring Ajay Dankar of early-SDN adopter eBay/PayPal ( EBAY ), Sean Roberts of OpenStack Foundation/VMware ( VMW ) and David Byte of SUSE Linux - recently acquired by MicroFocus (MCRO.L/[[OTCPK:MCFUF]]). Ajay Dankar, and his colleague Vinay Bannai in a prior panel, showed how eBay is employing automated virtual switching into its network today, as a beacon of hope for its slower-moving telecom brethren. eBay/PayPal has Infrastructure as a Service (IaaS) and Platform as a Service [PaaS] cloud solutions for some of its global operations. These cloud-based services are highly automated and simple to use - choose an application from a portal, and just click and run. So if there is a security breach, eBay/PayPal can flex-up or flex-down in a matter of hours via automated code to reissue IDs and other measures to resolve problems. This exemplifies the Agile network, as opposed to the historical Waterfall carrier protocol approach that takes several months via embedding, compiling, testing, etc. In fact, companies like Facebook ( FB ), Google, and LinkedIn (LNKD) run two production pushes per day. In a similar vein, David Byte of SUSE Linux explained how SUSE Studio can build cloud-ready applications that can be deployed with a push of a button. Lastly, Sean Roberts discussed the importance of setting common policy standards for Network Virtualization (of which OpenStack is a subset of).
OpenStack is open-source cloud-computing software. Open-source software is a key to successfully implementing automated networks that can scale and interoperate globally. Every six months, the OpenStack developers meet to improve the existing API (application programming interfaces). The carriers cry that OpenStack software is not perfect. But they are missing the boat on the many applications and functions that do not require 99.99999% accuracy. Software programmers pace of development dwarf the limited, snail-pace work of OSS/BSS carrier protocol specialists. Companies like SUSE can bridge the gap for carriers to deploy open-source Linux applications, in case they do not have the internal expertise like Google. This is the way that carriers can start to realize the agility/time-to-market, capex and opex savings, and reduced vendor lock-in that open-source cloud computing software can provide.
In the carriers' defense, they have complex, heterogeneous networks with high service requirements for legacy voice traffic, as opposed to the homogeneous internal networks of Web 2.0 data centers like eBay/PayPal. The problem is that operators like Century Link (CTL) are creating these more nimble networks that will challenge tier-one carriers in the future. Large carriers are aware of this threat, and are now starting to deploy SDN and virtualization on a small scale despite the hassles.
NTT DOCOMO (DCM) has taken the lead announcing successful NFV trials in October 2014 that integrated the software and equipment of six world-leading carrier vendors - Alcatel-Lucent (ALU), Cisco (CSCO), Ericsson (ERIC), Huawei (002502.SZSE), NEC (6701.TSE/[[OTC:NIPNF]]), and Nokia Networks (NOK). Using the virtualized Evolved Packet Core ( [[EPC]]), the core network of LTE mobile systems, the trials successfully verified the interoperability of each vendor's EPC software with a different vendor's equipment, such as network recovery in the event of hardware failure. NTT DOCOMO aims to utilize a virtualized mobile network for the commercial deployment of services by the fiscal year ending March 2016.
A key ingredient to NTT DOCOMO's initiative is the Open Platform for NFV (OPNFV) , a recently formed open-source project that brings together world-leading mobile operators and vendors to support development of an open-source platform for virtualization. Marc Cohn of Ciena (CIEN), a chairperson of OPNFV, discussed the benefits of OPNFV at the conference.
The AT&T (T) Domain 2.0 initiative includes a Virtualized EPC as well. Pari Bajpay of AT&T spoke about this new cloud-based design that incorporated open systems, AT&T APIs, customer cloud, and other components. While admitting that some gaps in the system like orchestration need to be addressed, Pari was positive on the long-term returns that network virtualization will provide. Pari projected that over the next five years, a significant portion of the AT&T network will be virtualized. Today, however, it is very early days in terms of actual virtualization deployments in AT&T's network.
In fact, the majority of most Web 2.0 data centers networks, as well as those of carriers, remain to be virtualized today. SDN and NFV require a business case or proof of concept that justifies the cost and risk of implementing the technology. Downtime for core activities like customer transactions at PayPal, as well as voice calls for AT&T, cannot be left to even the slightest vagaries that may exist in SDN or NFV. Thus, functions become virtualized over time from non-core to core as the technology matures, and as new business cases justify the investment.
During a roundtable discussion at our Palo Alto conference, I asked participants what initial applications made sense for carriers? One interesting answer was medical applications that can be automatically distributed over a mobile network. For example, patient heart rate and other readings can be monitored and sent to caregivers and doctors over a mobile network.
Containerization is a sub-technology trend that has been picking up steam this year. Containers are an automated allocation of workloads, which are replacing virtual machines (VMs) for running cloud applications. Containers require less memory and overhead than VMs, and can start applications more quickly. Amazon Web Services already offers Containers as a Service (CoaaS) on a pay per use basis. Not coincidentally, VMware's stock price has been weak recently, partially due to fears of containers replacing VMware's VMs. Previously, VMware (and its Nicira division) were perceived as an SDN beneficiary by investors. It is interesting to note that IBM (IBM) recently named containerization as one of its four drivers to generate long-term top-line growth, along with cognitive computing, security, and mobility.
Disclosure: The author has no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. The author wrote this article themselves, and it expresses their own opinions. The author is not receiving compensation for it. The author has no business relationship with any company whose stock is mentioned in this article.
See also Everything's Fine. Just Keep Investing! on seekingalpha.com
The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.
The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.
