It doesn’t seem that long ago that a colleague at diginomica said to me, as I pondered writing a piece about a potential new use case for Blockchain technology, ‘Don’t bother with that until it can actually do something that is proven to be of value’.
I could see his point, for while it provided some potentially powerful – and valuable – capabilities (not least being the ability to give provenance of ownership and/or contribution to an item or project of great value, such as jewellery, paintings – and yes, Bitcoins), it had the horrendous downside that wiped out its value for all types of trading except those of small volume/very high unit value.
The problem, of course, is that every node in a Blockchain has to be informed about every transaction in the chain, creating an exponential explosion of data generation – with all the subsequent storage management issues that automatically follow. It may create a secure record of any one Bitcoin, but the data explosion has led to protracted arguments about the energy consumption and carbon footprint of Blockchain.
This is an important point in terms of Blockchain’s value, particularly once the controversies around Bitcoin and cryptocurrencies are set aside. The transaction volume for all businesses using the internet to sell to customers and communicate with business partners and service providers is only going to rocket, particularly as cloud services extend out to incorporate everything happening out at the edge.
For example, the credit card giant Visa is reckoned to be able to handle some 1,700 Transactions Per Second (TPS), while in 2020, the Chinese online sales masters at Alibaba cranked up an amazing peak order rate of 583,000 per second during its now famous Single’s Day promotion. Now think about those transactions and what each of them means in terms of multiple business partners accessing processes that have, right out to the edge, thousands of data sources and associated process steps across several tens of closely collaborating partner companies, going all the way to the back office and all running on maybe hundreds of different machine types, software environments and applications.
This is a use case where Blockchain and its capabilities with transaction security and provenance could offer significant advantages, except for that one major deficiency, typified by the Bitcoin example: a TPS ranging between three and seven, and a transaction confirmation cycle time of around 10 minutes. In short, Blockchain can’t hack the business need or get remotely close…or can it?
OK, so that was the bad side
Vendia is a California-based start-up that aims to short-circuit this conundrum. It was co-founded in 2020 by Shruthi Rao and Tim Wagner, who had previously worked together at Amazon AWS, running its Blockchain business, and before that, its serverless business. Experience from both of those operations have gone into the philosophy underpinning Vendia, as Rao, now the company’s Chief Business Officer, points out:
When we ran the serverless business AWS customers loved it, but one main objection they had was that it doesn't have a data model. It's all compute. Another one was that it's a walled garden, AWS only, and big companies have AWS, but they also have some Azure, some on-prem, and they also have Snowflakes.
This customer reality then melded with input from running the Blockchain business, where customers kept observing that their need was to break down the walls that still existed between data silos. This was becoming a serious problem because most companies now work with partners, each creating ever more of their own data, but with no real means of sharing data between them, securely and appropriately, in real time. The hope was that blockchain might solve this problem.
Around this time, Amazon did an analysis which showed that 80% of a company's data is outside of that company's direct control. So the need is to make sure the data the business relies on is accurate, available, secure, and also available to/from partners. This is not made easier by partners having different types of tech stack, governance practices, operational policies and the like. Rao explains:
The most valuable businesses are data businesses, that's why we brought together the thesis of serverless and the thesis of blockchain, the immutability, the verifiability, the governance, and marry them together to build Vendia, which is short for Venn diagram.
For Blockchain, read distributed ledger
The result steps back from Blockchain proper and is instead based on distributed ledgers, the concept of multiple parties across multiple clouds, regions, and accounts having the same view of data, all in real time and with all of the semantics of a traditional database. But it operates in a serverless implementation, making it easy to spin up new partners, new regions, new nodes or new clouds without having to consider infrastructure. All that is needed is a schema, almost any schema.
Users get the whole tech stack including compute, storage, database, and networking for multiple parties in multiple accounts or multiple clouds. This includes taking care of asset semantics, multi-primary synchronization.
It also provides comprehensive access control to ensure that partners only see data appropriate to their contribution and the other side of that coin – the ability to apportion data so that appropriate partners can participate in delivering favoured customers specific types of deal or discount in real time, as Rao outlines:
An airline wants to give better rates to a particular customer company, not just through the airline, but on all of the aggregators such as Kayak and Expedia. There is no way to say an individual from this particular company gets a particular discount, irrespective of which aggregator the ticket is bought from.
This is all about being able to create smart contracts and, as she pointed out, you can’t do that if it takes a week to filter through that an individual from that customer bought a ticket. The airline obviously needs to know who is flying before take-off, so there is enough time for asynchronous transactions to take place. In the case of exchange trading execution needs to be almost instantaneous, and she claimed that all Vendia executions are completed in under 50 milliseconds.
According to Rao, the key to overcoming the Blockchain/shared ledger issue of exponential data generation is trading off a small amount of centralisation in order to gain a large benefit in performance and energy-efficiency. Vendia's consensus protocol ensures that traffic and energy use scales roughly linearly with the number of nodes in the system and is not prone to combinatorial explosion like traditional blockchain systems.
Unlike proof of work Blockchain solutions, Vendia uses a `low-trust’ proprietary cryptographic verification protocol coupled with serverless compute hosted in cloud service provider data centers offering high utilisation and efficiency. In this way, read requests in Vendia interact with a single node and do not consume resources on other nodes.
The BMW use case
An interesting example of the complexity the Vendia approach is targeting can be seen in a joint development underway with BMW. Like most car makers, the firm wants car owners to use its own branded network for communication, data transfer and the like, but the cost of establishing that infrastructure from scratch, worldwide, is horrendous, so third party services need to be used. The worldwide partner community is a key part of this. Should a BMW break down in the boonies of New Mexico on i-40 the car owner will want to see a large number of related actions from many different service providers being executed quickly and efficiently. Rao says:
They need to work with partners that may not be very tech-savvy, or may not have a technology competency or stack, even may not have another cloud NIC, even in AWS. And they have to work through all of that without worrying about another partner’s infrastructure. So the infrastructure layer has to be completely abstracted from them. They have to be able to define rules that are not just important for a business reason, but also for many different compliance needs, which will vary with location. And all of that needs to be considered in real time.
A final but important part of Vendia’s development is a workflow tool. It is interesting that the company has gone to the effort of developing its own from scratch rather than partner with an existing provider, but Rao says there is a specific reason:
This is more from a perspective of how to make it easy to build multi-party applications, rather than use something that focuses on building for one party, one account, on my cloud. You come with a worldview that is supposed to be looked like a certain way, but when you start doing multi-party applications, how do you test it?
This then plays to what Rao sees as Vendia’s primary objective, getting customers to obtain value as fast as possible, such as a production grade pilot within a week:
Four hours is all they spend with us. It's been very gratifying to see that you can do things fast. For God’s sakes, proof of concept doesn't have to be a year long. We don't even call it a proof of concept; we call it a proof of value.
For more years than I care to remember the technologies of IT infrastructure have been the key focus of everyone’s attention, often far more so than the reason for technology to be developed. The arrival of, first of all, the cloud, and now edge, the importance of technology and technology differentiation is diminishing fast. Indeed, technologies that get in the way of collaboration and where differentiation becomes a specific barrier across the widest possible range of different technologies must now die out. By the same token, new technologies must find the way to muscle in, especially if they something to the party.
It can be argued that Vendia does that, in at least two ways: one is that it specifically targets multi-cloud environments agnostically. To do otherwise is to provide roadblocks to the sharing of data in a world that now needs to have data shared. The other is that, by taking a step back from the Blockchain archetype and focussing on distributed ledgers (and avoiding calling it ‘Blockchain Lite’), users get the advantage of secure shared data at scalable speeds of operation.