How Disaggregation is Paving our Path Forward

Abstract background from metallic cubes

The optical networking industry is on the edge of revolutionary change, and while it may sound trite to talk about “the network of the future,” this is what’s approaching in a very real and immediate sense.

Two long-term market trends—industry consolidation and the convergence of IT and networking—have propelled the industry to its first major inflection point in decades. Two technological trends—the software revolution and the disaggregation phenomenon— are taking us forward.

Of these trends, disaggregation is the key to open, agile, plug-and-play networking as we will know it in five years’ time. This is because before we can progress, we have to separate the individual functions and capabilities that comprise today’s tightly integrated hardware systems. Disaggregation is, therefore, not just the latest buzzword. It’s a prerequisite that must be met before we can form an ecosystem of new industry partnerships, and collaborate to rebuild and improve those componentized functions—and add the automation and intelligence that the market’s clamoring to buy.

Fujitsu is shifting the architectural design of its optical networking platforms away from complex, vertically integrated hardware-based structures into a disaggregated architecture. This change will give rise to re-aggregation in the form of economical, generic hardware elements, open-standards software frameworks, and interoperable functional modules. We call this “componentization.”

Once disaggregation has happened, we can drive fast implementation times, simpler development cycles, lower costs, and an overall climate of unprecedented innovation, collaboration, and opportunity not just for the industry itself, but for our customers and their customers as well.

Innovation, collaboration and opportunity—the broad benefits resulting from the disaggregation journey—all grow best in an open, nonproprietary climate. They thrive when the overall environment is flexible, partnership-oriented and fast-paced. And while we’re aware of and prepared for the inevitable transition period of backwards compatibility, we’re committed to taking the first bold steps down the road to disaggregation in fall 2015. It’s an exciting time to be at Fujitsu, an exciting time to be in the optical networking business, and it’s going to get more and more interesting from here.

The Promise of VoLTE: It’s Only a Matter of Time


Voice over LTE (VoLTE) is considered by many to be revolutionary both for mobile operators and for subscribers. Operators, once they have established their VoLTE networks, will no longer have to maintain separate networks for voice (circuit-switched) and data (packet-switched). This will save on operational and capital expenses. Subscribers who use VoLTE will be able to use high-quality voice and data applications simultaneously, and the clarity of their voice calls will improve.

So why has VoLTE taken longer than anticipated to deploy? The answer lies in several challenges, which I’d like to discuss briefly.

  • The successful roll-out of HD voice and video calling services requires VoLTE technology to be simultaneously available in both the mobile network core and on mobile handsets. Most mobile operators’ core infrastructures are not fully equipped to simultaneously support circuit-switched and packet voice. In addition, VoLTE-enabled handsets are not yet widely available.
  • VoLTE promises to move wireless calls from the legacy circuit-switched network to the all-IP-based LTE network. The formidable task of supporting both switched-circuit and packet-based technologies, however, is not economical for mobile operators in markets where LTE is not yet deployed. What’s more, mobile operators are still resolving VoLTE call interoperability issues to support customers who are roaming.
  • Finally, successful VoLTE deployment depends heavily on nationwide LTE deployment and the adoption of LTE-based small cells for in-building voice enhancements. Adoption of LTE-based small cells for residential and enterprise applications is still very low.

Despite these challenges, numerous efforts are underway to realize the promise of VoLTE.  To expedite nationwide rollout, several leading mobile operators have begun limited trials in major markets, testing voice quality and equipment interoperability. Suppliers are slowly rolling out VoLTE-enabled handsets; and a consortium of mobile operators is negotiating roaming agreements for smooth inter-network transitions. The potential long-term benefits of VoLTE to operators and consumers alike are too great to miss and they easily outweigh any temporary intermediate setbacks.

Real-World SDN: Anything Less than Multivendor Won’t Cut It

Abstract background from metallic cubes
Vendors seeking to set “ground rules” for critical aspects of Software-Defined Networking (SDN), such as what “multilayer” approaches mean, are leaving gaping holes in their strategy if they don’t account for multivendor interoperability. A diverse vendor ecosystem is one of the things that makes an open standards paradigm so powerful.

The importance of multilayer networking is beyond dispute, regardless of what the “routers can do everything” lobby might have to say. In fact, the idea that entire networks can be based on router architectures amounts to a case in point for the multilayer view. Trying to build an end-to-end router-based network would impose tremendous cost and complexity burdens. The cheaper, more efficient, long-range optical transport layer is essential to economical networks, alongside routers. It’s a question of using the most efficient means to carry traffic, not taking a one-size-fits-all approach.

In a similar vein, multilayer approaches seem to be coalescing along single-vendor lines, or at least to be offering limited multivendor interoperability. Even those claiming to be closest to rolling out a multilayer SDN offering seem to imagine that service providers will buy all their equipment from them. When the chips are down, multilayer approaches based on alliances between vendors to make up for each other’s deficiencies amount to turf-protection efforts that don’t take account of the diversity at all layers of the network. In reality, a truly open and interoperable multivendor approach ranks alongside multilayer networking in importance. Yet we hear very little about the multivendor aspect save for the claims of one or two vendors of glorified network management systems thinly disguised as SDN offerings.

Any single-vendor or vendor-limited solution is useless in the real world where SDN will be deployed. Moreover, protecting old OSS turf prevents interoperable solutions from flourishing and produces some “strange bedfellows.” A realistic and pragmatic approach to SDN must recognize the true conditions into which SDN will be deployed. In a fragmented optical market and an economic climate where service providers seek to maximize the long term viability of their capital investments, SDN must be as interoperable as possible if it is to deliver on its value promises.

As the provider of one of the first multivendor SDN solutions deployed in the world, Fujitsu has valuable expertise and perspective on this topic and is uniquely positioned to address connecting data centers to transport networks using multilayer and multivendor systems. We already build servers and resource orchestrators, and are currently among the world’s largest providers of hosted cloud services. Look for more announcements from Fujitsu on their fast-developing SDN technology portfolio over the coming months.