Coindesk’s recent article "Banks taking sides"  covers the current race to market for competing trade finance initiatives and makes many valid observations on the current situation but also points to some key questions.
Just to summarise, there are more blockchain initiatives underway in Trade Finance than pretty much any other area of financial services. They may be characterised by their choice of technology (most commonly either Hyperledger Fabric or R3 Corda), by their business focus (open account or the documentary credit side), or by their geographical focus.
 
 So we have:
 We.Trade (HSBC, Deutsche, Rabo, KBC, Nordea, Santander, SocGen, Natixis, Unicredit) who are focused on SME trade finance, in Europe, using Hyperledger Fabric, and is the first of all the competing initiatives to actually be in production (as opposed to pilot or PoC).
 
 Batavia (UBS, Erste, Commerz, BMO, Caixa) with a broader business scope and geographical focus again using Hyperledger Fabric.
 
 Marco Polo (according to Coindesk formed of NatWest, BNP Paribas, Commerzbank, ING, Standard Chartered Bank, Natixis, Bangkok Bank, SMBC, DNB and OP Financial Group) with again a broader business and geographical focus, using an application from TradeIX, and R3 Corda as the underlying technology 
 
 Voltron: HSBC and other banks, focused on the documentary credit side, also using R3 Corda
 
 HTFP — led by HKMA, with HSBC, Standard Chartered Bank, Bank of East Asia, ANZ, Hang Seng and DBS Bank, focused on trade from Hong Kong, using an modified form of Hyperledger Fabric, to be implemented by Ping An.
 
 GTCN — the effort to connect Hong Kong and Singapore trade, working with reportedly 20 banks, yet to confirm a technology decision.
 
 .. and others including some not yet in the public domain.
 
 This raises three key questions as touched on by Coindesk:
 
Balancing Time to Market against Ambition
To answer this, we might recall General Patton’s quote ““A good plan, violently executed now, is better than a perfect plan next week”. In the commercial context, time to market is critical, and as every fintech knows, it really is the case that Good/Now trumps Perfect/Later. Good/Now means clients committing to a platform, and with the right “stickiness’ in the design and business proposition, the opportunity to leverage that famous first mover advantage.
 Time will tell but We.Trade in my view has reached a critical milestone in being in production, not just for time to market advantages, but for having cracked the code for what running a distributed private/permissioned network in production really means in banking.
 
So what’s the big deal about being in production?
Blockchain purists argue for a completely distributed model with no central roles or functions. However in the trade finance area, where banks are regulated, clients must be KYC’d, and business networks need to agree the rules by which they will operate, this model does not apply. 
 In this context setting up a trade finance network and running pilot transactions is "relatively" easy. Agreeing the compliance, security, governance, operational rules etc between a significant number of competing banks is anything but. Rules need to be agreed on how new banks will be on-boarded, technical models need to be developed to facilitate this easily. How smart contract code across a network can be updated needs to be agreed, and out of region Disaster Recovery processes defined. The translates into a major effort in defining and agreeing the components of the network’s operational and service delivery model. A large number of participating banks means potentially unmanageable numbers of bank representatives involved in this detailed planning, which is why WeTrade’s establishment of a JV entity has greatly simplified the path to production. As in the words of Roberto Mancone the WeTrade COO “We are ahead of the pack because we are not a consortium.”
 
And finally, interoperability
As noted by Coindesk, so many competing initiatives raises the question of whether trade finance clients are well served by an increasing number of potentially siloed networks, so integration of separate networks or delivering interoperability becomes a key topic.
 However I have to take issue with some of the comments made in the article; it is clearly beneficial that Corda can facilitate interoperability between Corda-based networks such as Marco Polo and Voltron; but I dont think anyone can claim true interoperability if it is dependent on a single choice of technology — be it Corda or anything else. Email (SMTP) interoperates fine without requiring us all to be using Outlook. And we know one blockchain technology wont rule them all.

Interoperability should also be seen in the much broader sense than the pure blockchain technology — ie the corporate governance model of a network, its operational rule book, data, and all the other non-blockchain components of the platform. Our experience of platforms like WeTrade is that only ~25% of them are the blockchain components. On this basis one could argue that all networks are in effect proprietary at the moment, in practice.
 So yes, interoperability is important, but its a bigger question than DLT’s talking to each others, and in my view is best served by resolving within the open-source, open-governance Hyperledger community, where after all many of the largest banks, corporates, R3, and IBM are all members, overseeing multiple technology projects including Fabric and ethereum-based ones.
 
 So in summary, banks face many choices — and not necessarily exclusive ones, eagle-eyed readers will notice some banks appear in multiple initiatives; bets are being hedged, and the race is most definitely on. The goal of being in production should not be under-estimated in terms of complexity. The operational model in this environment is complex, and there needs to complete clarity on which actor (client, bank, application provider, blockchain provider, systems integrator etc) is responsible for which element of the service delivery. Blockchain RACI models are vital to flush this complexity out.
 
 And of course, the banking industry — and more importantly bank clients, will be not be best served by a fragmented set of functional and technology silos, so consolidation and/or interoperability in the broader sense need to be considered to avoid a patchwork quilt (no hyperledger pun intended) of networks.