YOU ARE AT:Opinion2015 Predictions: Security for IoT; orchestrating NFV, SDN prime for progress

2015 Predictions: Security for IoT; orchestrating NFV, SDN prime for progress

Editor’s Note: With 2015 now upon us, RCR Wireless News has gathered predictions from leading industry analysts and executives on what they expect to see in the new year.

Security leads off 2015 predictions

As 2015 gets underway, it’s natural to consider what’s next on the horizon. It’s obvious that wired and wireless connections continue to grow, particularly as previously unconnected devices and formerly “closed” networks join the broader wide area networks. But what else can we anticipate in conjunction with this growth? Here are my predictions for 2015.

Security is the linchpin to success of “Internet of Things”

We’re confronted almost daily with news reports of security breaches. As the number of connected “things” continues to explode, we’ll only see more challenges in 2015. Why? These threats are no longer limited to consumer identity or credit card theft, but now also target industrial IoT networks. We gain the largest economic benefits by connecting more and more systems to each other and the cloud, but this very connectedness also makes security more paramount than ever. Since there’s no one security scheme to address all possible vulnerabilities, securing applications, networks and devices is imperative. Authentication, authorization and accounting are applicable to all these aspects and, combined with data confidentiality (encryption), are now widely used for the applications layer. Unfortunately, application security means little if the same AAA principles are not enforced for the network and device/link layers as well. The latter may be especially vulnerable, since street-level cells could easily be swapped out with rogue devices, which could then access the entire network. Fortunately, thanks to the availability of link layer AAA and encryption technologies like IEEE 802.1AE with strong 256-bit encryption, those last-mile links can be secured.

Management and network orchestration will be pivotal to the SDN/NFV network revolution

Software-defined networking and network functions virtualization are often linked, but in reality are quite different. What they have in common is that both look at the network within the context of an IT-centric model, vs. the traditional operational network view. Abstraction of network and compute resources from the basic network and server infrastructure is the IT-centric model’s chief advantage. This then enables those resources to be “orchestrated” to quickly deliver services to customers within hours or minutes vs. months in today’s model.

We see the management and network orchestration aspect, and to a lesser extent, the networking and compute infrastructure, as the largest opportunity for the industry to revolutionize the network. Revamping organizational skill sets and structures to adopt the new IT-centric model will make this an arduous and profound shift for network operators. However, doing so will pave the way for the SDN/NFV network revolution.

SDN will take its toll on switch and router sales in the short-term

There’s been plenty of industry focus on SDN and NFV, which will one day transform networks globally. However, large enterprises and network operators are currently faced with deciphering what SDN/NFV means to them, the business opportunities and how to evolve their networks to a model that’s more software- and IT-centric. This will hinder switch and router sales in the short-term.

SDN is a challenging value proposition especially for wide area network operators. While SDN’s flexibility can potentially decrease operating expense and capital expense, because most network equipment is already in place and won’t be fully amortized for many years, it’s considered “free.” Figuring out how the existing network can function within the IT-centric operational model – with its more centralized control planes and networkwide orchestration – will be the key to SDN’s success. Switch and router sales will strengthen again once that’s resolved.

Ethernet gets deterministic … and industrial IoT goes all-IP

Although Ethernet originated as a local area network technology over 40 years ago, it now serves the “language” to deliver high-bandwidth services in all kinds of networks, replacing semi-proprietary and legacy networking protocols. And it’s about to happen again – in industrial IoT.

Enterprise LANs were among the first to migrate to Ethernet in the late 1990s, with carrier networks following a similar path a decade later, replacing protocols like SONET/SDH, PDH, frame relay and ATM with Ethernet. Just as Ethernet came to dominate in these broad-based networks, we see the beginning of a similar transition within industrial IoT. Over time, industrial IoT networks will expand, connecting billions of “things” in both consumer and industrial applications. As with other networks, standards-based Ethernet will replace the legacy industrial IoT protocols such as Profinet, ControlNet, Fieldbus and Hart. The reason why is simple: Ethernet standards and protocols have evolved to handle virtually all of the industrial market’s particular requirements such as network synchronization, high availability and quality-of-service.

This year will bring the last piece of the puzzle to Ethernet that industrial IoT networks need: determinism. Making Ethernet deterministic translates to strict time windows within which packet delivery and acceptance by mission-critical devices can be guaranteed. With the IEEE 802.1 time-sensitive networking standardization effort, Ethernet switches will precisely know network time and be able to deliver queued packets either within a certain time frame or at a specific time. Deterministic Ethernet technology will be highly relevant to real-time industrial process control and factory automation systems. It also will be significant to other real-time applications, including transportation control and vehicular networks with mission-critical systems such as collision avoidance, advanced driver assist and antilock brakes. We expect the automotive market to be the next domino to fall in the network migration to Ethernet.

Connected car: Ethernet connections in automobiles will surpass the number of all other Ethernet-connected devices globally by 2020

One of the largest electronics markets is automotive, where the trend is toward the connected car. Every leading automotive manufacturer now supports Ethernet as the best alternative to the semi-proprietary and legacy protocols used within cars to accommodate the increasingly complex network requirements for bandwidth and connectivity. Connecting things like backup cameras and infotainment systems via Ethernet was just the start. We’ll soon see Ethernet as the backbone for highly critical systems including advanced driver assist and collision avoidance systems, diagnostics and ultimately the car’s main system bus.

Analysts expect that over 120 million cars will have Ethernet connectivity by 2020. So-called “premium” cars will connect up to 35 systems over Ethernet, with “mid-range” cars connecting between eight and 20 systems. This calculates to approximately 500 million to 600 million ports, which exceeds the number of Ethernet LAN ports today. Key industry initiatives enabling this: the IEEE 802.1tsn “Deterministic Ethernet” standardization and the new two-wire Ethernet PHYs coming out of the IEEE 802.3 100Base-T1 and 1000Base-T1 standards efforts.

Vitesse_MNuss_093_5x7

Dr. Martin Nuss is VP of technology and strategy and CTO at Vitesse Semiconductor. Nuss has over 25 years of technical and management experience and is a recognized industry expert in Ethernet technology including timing and synchronization for public and private communications networks. He serves on the board of directors for the Alliance for Telecommunications Industry Solutions and is a fellow of the Optical Society of America and IEEE member. He holds a doctorate in applied physics from the Technical University in Munich, Germany.

ABOUT AUTHOR