Hacker News Reader

Comment by OJFord | Hacker News Reader

Goes to show, you can shake-up a sector with your modern infrastructure all you want, if you still have to plug into it somewhere...

To be completely fair to Monzo though, my limited understanding is that this reliance on a 'supplier' for payments will go away when the beta ends and Monzo's a full-fledged bank/current account, rather than just a prepaid card.

EDIT: ^ That was correct, that is the intention: [0]

> When we launch the current account later this year, it will not use a third-party card processor and instead transactions will be processed entirely on our own technology.

[0] - https://monzo.com/blog/2017/03/05/outage/

View on HN

They're currently dependent on Wirecard for the underlying processing until they get their full banking license. At which point they can ditch that dependency - that's expected to be in ~May this year.

Edit: My mistake, its GPS doing the processing -- who are currently experiencing issues.

So it is the Wirecard infra failing then? And it depends if they can ditch that dependency; they will have to work with GPS or whoever themselves then. That is quite a lot of work. There are companies providing an intermediary to lessen that burden. Ideally you would want to cut it out indeed.

I believe it is specifically GPS, however I believe they are also doing that directly for their current account launch - they're putting a lot of emphasis on doing everything in-house

> I believe they are also doing that directly for their current account launch

They are, I just updated my top-level comment - they confirm that in the blog post the CTO linked in another comment here: https://monzo.com/blog/2017/03/05/outage/

Wirecard are the issuing bank but not the card processor.

the processor is the dubai based GPS https://twitter.com/GPS_Processor

My crystal bubble says it's gonna be a fun time when they'll roll the current accounts and be the payment processor.

It seems that currently they are (mostly) stateless and build around statelessness. i.e. get clients orders and pass it back to the payment providers.

Once they have current accounts, they'll have to hold serious funds and will be held to a much higher standard, which is a whole level of new difficulty to achieve. (unless they find a 3rd party service to outsource all of that :D).

Source: Switching from running market-makers to financial exchanges. Same stuff behind the hood, one is 10 times harder than the other one.

But then we can just have Monzo's own infrastructure fail or so. In saying this, I'm a happy Monzo user for almost a year and use it regularly, but it really dawned on me today that new technology brings new ways for things to fail, and when it's to do with my bank, it worries me a bit.

The nature of the way Monzo operates makes this much less likely to occur.

And if it does happen, the nature of how they operate means it should be significantly faster to resolve. Particularly because there is no reliance on a third party

Pedantically, they'd still be reliant on MasterCard's systems, and their connectivity to them, but if MasterCard went down Monzo wouldn't be the only one with problems.

Care to explain why? What is this "nature of monzo operation" that solves errors quicker?

If anything, a small startups who has to redo everything from scratch with limited resources seems like there is huge room for errors, and not many people to support and fix them 24/7.

They are a tech company first, financial one second. They made a decision very early on to build everything from scratch in the "right" way, pretty much no matter how long it would take or how much effort would be involved.

Their systems are incredibly stable. In fact, I don't think I can recall a single bit of downtime which was actually down to their system since November 2015 when I joined.

In the case which happened here, some kind of unexpected migration gone wrong - Monzo's systems are highly decoupled and separated. So any given database issue will only affect that given service. As each service is so simple and contains relatively little code, working towards a resolution can be achieved significantly faster. The smaller scope of responsibility also means that a mistake is less likely to happen in the first place.

> Their systems are incredibly stable. In fact, I don't think I can recall a single bit of downtime ---which was actually down to their system--- since November 2015 when I joined.

Downtime is downtime.

Ignoring and shifting the blame to third parties will not fly with customers who are unable to buy their groceries. Just saying ;)

That is exactly why they're bringing it in-house so the blame would be on them and they're in control.

Reply on news.ycombinator.com