DCI BOX is also called modular wave division device. In fact, DCI BOX does not require much technical precipitation, because the most important core chip/optical module can be purchased. So why was DCI BOX created? In my opinion, it mainly depends on the scene and the cost. In the scene, DCI BOX is the architecture for the next generation MAN, such as metropolitan distribution (POD), cloud network access point (POP) and other functional areas. Take the Spine-Leaf structure of POD as an example. The flat structure of this new MAN has high requirements on service delay, so the point-to-point networking is generally adopted.
Data centers also have high requirements on link bandwidth, which leads to a large demand for lateral expansion of Spine-Leaf architecture. Especially in the computer room space, the traditional WDM/OTN frame type large equipment is difficult to cope with this explosive growth. Generally, metro DC interconnection is a point-to-point scenario of less than 300km, which does not have large requirements on performance such as transmission distance. Therefore, large devices are not required.
The above is the technical scenario aspect. Back to the most essential issue, it is still the cost consideration. Those who do WDM projects generally know that the unit price of WDM equipment is very expensive, so operators or OTT vendors do not want to be tied to a single manufacturer in terms of price. Therefore, as with organizations such as Open ROADM, operators and OTT vendors are particularly interested in making standard customized gray box devices that are photoelectric decoupled and can be based on unified control (open), thus getting rid of the deep binding with specific vendors.
The networking of the DCI BOX application scenario is as follows. The optical layer and electrical layer devices can be networked by different manufacturers.
Advantages: small size, easy to expand, and quick business opening.
Next, let’s briefly talk about two more sensitive issues.
The first problem is open decoupling. Both traditional large equipment and DCI BOX here cannot achieve complete decoupling at present. It remains to be seen whether open-Roadm, an Open organization related to wave division, can successfully promote its standards. That is to say, at the most, the device shape, size, port identification and so on should be unified, including optoelectronic equipment level decoupling: the OTU service board of manufacturer A can connect to the OTU service board of manufacturer A through the optical layer (splicing and combining wave) of manufacturer Z. In fact, this is similar to the external wavelength scheme on the traditional wavelength division OTN equipment.
However, due to the differences in internal chip, backplane, line framing and FEC coding, it is not possible to make the OTU service veneer of manufacturer A compatible with the equipment slot of manufacturer Z, or to connect the OTU service veneer of manufacturer A to the OTU service veneer of manufacturer B.
In addition, in terms of operation and maintenance, multiple manufacturers are mixed network. Each manufacturer can cooperate with the operator to develop customized software based on the open northbound interface standard. However, this will certainly increase the difficulty and complexity of operation and maintenance, and the positioning of the fault interface is difficult to grasp.
The second question is about cost. Here it is necessary to briefly mention the progress of major chips/devices in China’s localization, i.e. the current supply chain situation.
In terms of optical layer core devices, WSS devices for ROADM architecture and passive co-splitwave devices have been gradually developed by several traditional Chinese equipment manufacturers and devices, while other manufacturers are most likely to purchase devices such as Cohrent.
In the case of electrical layer devices, the situation is basically similar. For example, DSP chips and OTN frame chips in high-rate line-side coherent optical modules may still be mainly purchased by small manufacturers, and the supply chain is relatively simple. It is said that at present, only Microchip frame chips are sold out of OTN, and DSP chips can be provided by European and American companies.
Therefore, China’s domestic operators and OTT manufacturers want to increase the proportion of core devices and strengthen the localization of China, firstly for supply chain security, and secondly for reducing the cost of self-developed equipment, which requires a long-term upstream and downstream cultivation of the industrial chain.