data disaggregation image credits

How about programmable functional data disaggregation?

November 15, 2016


Programmable functional disaggregation is the new buzzword in the telecom industry. Telco Transformation mentions this type of data disaggregation in one of its September editorials. It involves breaking up network equipment into smaller, more flexible and leverage components. Basically, this concept allows solutions architects to assemble custom, open source elements into personalized suites.

Fujitsu Network Communications strongly supports this disaggregation movement. They see it as the answer to the current need for fundamental design changes. In their view, the networking industry could greatly benefit from programmable functional disaggregation. Their paper entitled “Moving towards the programmable, disaggregated network” (accessible here) illustrates the way this trend serves the new types of modern networks. Mostly, such recent networking solutions sprung from the phenomenon of border blurring between the IT and the Communication networks’ worlds.

Two levels of functional data disaggregation

As Fujitsu Networks’ Joe Mocerino, principal solutions architect of packet optical network points out, there are two levels of functional disaggregation.

The primary level consists of disconnecting the optics from the actual plug in models.  The purpose is to enable an entire suite of different optics.

The second, more in-depth level, concerns the open APIs and the cross-functional standards. Once disaggregated, these allow the clients to connect one vendor’s transponder to another vendor’s reconfigurable optical add-drop multiplexer (ROADM). This way, the entire transport layer can be suitably composed out of various pieces of equipment coming from different vendors. Functional disaggregation allows better scalability and more flexibility.

Benefits of programmable functional disaggregation

One may better visualize the entire functional disaggregation process as completing a puzzle. Of course, your preferred vendors supply the pieces. These must be compatible and easy to fit together. Just think how networking this way resembles to a construction made out of Legos.

As long as the matches are functional and benefit the network and its overall functionality, the cost savings can be significant this. Travis Ewert, (senior vice president of global network software development at Level 3 Communications Inc.) looks at how third-party pluggable optics generate roughly a $10 million a year savings. He estimates that fully programmable functional disaggregation would considerably increase the savings’ value.

Fujitsu considers that programmable functional disaggregation’s compelling benefits would manifest themselves in networks of all sizes. More exactly, they mean:

  • Efficient scaling that would easily adapt the network fluctuating size over time;
  • Efficient, space-saving rack space utilization;
  • Increased innovation, due to removal of dependency limits.Innovation is permitted once component design benefits from the horizon of independent functions and software;
  • A brave new world of open network architecture, where solution architects would be able to choose the “best–of-breed equipment”.

Experimenting with programmable functional disaggregation

We have located two online available papers on programmable disaggregated networks. In short, they both illustrate how this new concept stands in practice.

One paper reports “an FPGA-based switch and interface card (SIC) and its application scenario in an all-optical, programmable disaggregated data center network (DCN)”.  The other one presents “an all-optical flexible disaggregated flat DCN architecture utilizing Hollow-Core Bandgap Fibre, reconfigurable 4×16/8×12 Spectrum Selective Switches and FPGA-based switch and interface intra/inter-blade cards enabling chip-level access while utilizing cut-through low-latency tuneable WDM/WDM-TDM”.

Both examples are interesting to browse for the concerned specialists. Overall, they provide an anticipatory image of what programmable functional disaggregation could bring into the optical communications field and its network-related applications.

Reasons for skepticism in data disaggregation

It seems like the guarantee of cost savings is not as certain for some. Programmable functional disaggregation is just emerging. There is no certain telling yet whether putting it in practice on a real-life scale would not generate different figures.

Also smaller service providers might not be interested in the promises of network disaggregation, due to their scale of operations.

Network Computing explained in an early 2016 article why network disaggregation does not make sense for most enterprise networks. Their main arguments were the following:

– Lower costs are in fact a false promise in this case;

– The incumbent need for network platforms to be certified by the OS vendor in order to be compatible with the chosen OS;

– The lack of readiness for business when it comes to funding necessary solutions in order to automate network management with new switches deployment that needs puppet/REST API;

– The questionable simplification brought by reduced vendor lock-in;

– The questionable advantage of running other software on the switch in times when traditional vendors start to offer this option too.

A wind of change

Network Computing’s article is a reaction to a Forrester report entitled “The Myth of White-box switches”. The object of the debate therefore consists of switch-related types of network disaggregation structures. Nevertheless, such argumentation seems to nip in the bud the prospect of Lego-like networks, at least for some types of operations.

It is still early, though. The Internet giants might invest in this emerging innovative technology. As always, they set the bar when it comes to next-gen technologies. The industry has not decided yet in what network disaggregation is concerned. The wind of change might feel like a breeze for now, but its presence is not negligible.

Here you may see how Forbes juxtaposed the pro and con points of view on this matter, as early as 2014. Furthermore, the dispute is not only technology-related, but also market-related. The open source supporters slowly try to segment and open up certain sub-markets to the detriment of the traditional legacy vendors. Other industries replay the exact same move as well – a fact that tends to become a trend.