Data sharing in industrial ecosystems

Data sharing in industrial ecosystems (continuation)

by Prof. Dr. Boris Otto, Prof. Dr. Niko­laus Mohr, Mat­thi­as Rog­gen­dorf, PhD and Tobi­as Gug­gen­ber­ger

Lay­ing the ground­work for mul­ti­par­ty indus­tri­al data sharing

Com­pa­nies have been sharing data in bila­te­ral, trus­ting rela­ti­ons­hips for deca­des. For examp­le, they may tra­de infor­ma­ti­on in clo­sed sup­ply chains to make ven­dor-mana­ged invent­ories pos­si­ble or in the finan­ce space to auto­ma­te invoi­cing and pay­ment flows.

Howe­ver, an eco­sys­tem is far more com­pli­ca­ted. For examp­le:

  1. The data invol­ved is often more com­plex, and the par­ties do not align on data onto­lo­gy in advan­ce. If no indus­try stan­dard exists, par­ties may nego­tia­te pro­prie­ta­ry stan­dards.
  2. Mecha­nisms for ensu­ring data is used as agreed are much har­der to enforce and moni­tor.
  3. Value from data needs to be sha­red wit­hin the eco­sys­tem to keep par­ties inte­rested in con­tri­bu­ting. Imba­lan­ces quick­ly lead to dro­pouts that may end­an­ger the sta­bi­li­ty of busi­ness models.

On the posi­ti­ve side, oppor­tu­nities from indus­tri­al eco­sys­tems are sub­stan­ti­al­ly hig­her than for data-sharing set­ups that func­tion as one-trick ponies. As data sharing expands bey­ond sin­gle use cases and data usa­ge, par­ti­ci­pants may explo­re dif­fe­rent types of inno­va­ti­on. The­se may invol­ve blen­ding data with other sources as men­tio­ned above, app­ly­ing machi­ne lear­ning or arti­fi­ci­al intel­li­gence, or extrac­ting value from streams of real-time data that might dimi­nish as the data beco­mes sta­le. With the pro­li­fe­ra­ti­on of IoT devices and Indus­try 4.0 adop­ti­on, ever-gro­wing amounts of data – with lar­ge­ly unex­plo­red value in many cases – are beco­m­ing avail­ab­le.

Indus­tri­al data eco­sys­tems also play an important role in the suc­cess­ful use of digi­tal twins. A digi­tal twin is a vir­tu­al rep­re­sen­ta­ti­on of a machi­ne, a pro­duc­tion line, or even an ent­i­re pro­duc­tion envi­ron­ment such as a fac­to­ry or (as in semi­con­duc­tor or bat­te­ry pro­duc­tion) a net­work of fac­to­ries. While bila­te­ral data sharing might be suf­fi­ci­ent for imple­men­ting a digi­tal twin of a sin­gle machi­ne, exten­ding its scope requi­res a more advan­ced approach: a sha­red digi­tal twin invol­ves mul­ti­ple part­ners and sup­pliers, each with spe­ci­fic inte­rests, avail­ab­le data, and inter­nal capa­bi­li­ties for lever­aging and sha­re infor­ma­ti­on.

Tech­ni­cal solu­ti­ons are nee­ded for recei­ving, sto­ring, and sharing – and, depen­ding on eco­sys­tem poli­ci­es and rules, pro­ces­sing and ana­ly­zing – data from the ecosystem’s dif­fe­rent mem­bers. Many indus­tri­al com­pa­nies have exten­si­ve­ly con­si­de­red spe­cia­li­zed data and IoT plat­forms to help them mas­ter their data sharing chal­len­ges. And they have a wealth of opti­ons: in recent years, more than 1,000 such offe­rings have hit the mar­ket [1]. Howe­ver, many indus­try decisi­on makers feel that the­se plat­forms only par­ti­al­ly meet expec­ta­ti­ons. While they gene­ral­ly pro­vi­de a solid basis for hos­ting mas­si­ve data volu­mes or sho­wing real-time data strea­ming pat­terns, they fall short in terms of col­la­bo­ra­ti­on and sharing capa­bi­li­ties. Often, they are only desi­gned for sin­gle-par­ty use. Big cloud hyper-sca­lers often offer more fle­xi­ble plat­form ser­vices and allow com­pa­nies to sha­re data more open­ly while stay­ing in full con­trol of the­se data assets.

[1] VDMA: Platt­form­öko­no­mie im Maschi­nen­bau. Her­aus­for­de­run­gen – Chan­cen – Hand­lungs­op­tio­nen

To the next part.