Restarting the free accounting search

Posted Jul 31, 2017 13:07 UTC (Mon) by felixfix (subscriber, #242)
In reply to: Restarting the free accounting search by fujimotos
Parent article: Restarting the free accounting search

Worries about a vendor going out of business or raising their prices or changing anything should not be restricted to cloud services. What if the landlord goes bankrupt or doubles the rent or just plain doesn't want to renew the lease? What if the city executes eminent domain and throws you to the curb? What if the data center runs out of space or has a power failure or someone else's rack fries itself and takes out your rack? What if the garage that services your vehicles relocates or goes out of business? What if the hotel you have a deal with gets a better deal from a larger company, or a convention hits town and books all the rooms just when you need them? What if the roof air condition dies and it will require city permites to get a permit for the crane and several weeks for the parts?

There is nothing special about cloud vendors.

Restarting the free accounting search

Posted Jul 31, 2017 13:35 UTC (Mon) by pizza (subscriber, #46) [Link] (18 responses)

The key difference between "your cloud provider" going away and everything else you described is that none of the latter will result in your finances being held hostage.

If your landlord screws you over, you have volumes of laws and regulations to keep you able to at least negotiate while looking for other alternatives. The same for eminent domain seizure, which doesn't just happen overnight. If your roof A/C dies, you bring in portable units to keep your essential equipment cool for as long as it takes.

As for hotels, there are many out there, and conventions are known out a year ahead of time, making it trivial to plan around such things. Ditto for a service center; that's a true commodity.

But your accounting platform is _not_ something trivial to migrate, especially at the last second. That going away will literally put you out of business.

Restarting the free accounting search

Posted Jul 31, 2017 13:50 UTC (Mon) by felixfix (subscriber, #242) [Link] (17 responses)

Every company has a zillion weak points. Singling out cloud accounting data requires wearing blinders.

Saying there is legal recourse implies legal recourse is speedy. That's a laugh!

Saying conventions are known a year in advance ignores all sorts of complications which can happen and have happened. Fires, Legionnaire's Disease, bureaucratic snafus, anything can disrupt plans. Conventions may prove to be more popular than expected. The President may schedule a campaign speech or international political event and disrupt business as usual.

How many big name companies have forgotten to renew their domain names?

You rent five floors in a skyscraper. Where are you going to put those portable A/C units, which require sealed rooms and outside air? Where are you even going to get those portable A/C units when everyone else in the building also wants them? Maybe it will only take a few days -- how long can you afford to be shut down?

Who cares if your finances are being held hostage if your ISP gets knocked out by a hurricane? What if yoru internet connection is clobbered by a ship dragging anchor?

Finances are not the only, or even the most, critical part of a company. That's like saying tires are the most critical part of a vehicle.

Restarting the free accounting search

Posted Jul 31, 2017 14:24 UTC (Mon) by pizza (subscriber, #46) [Link] (4 responses)

I just went through a round of "business continuity" planning with my employer. They take this stuff very seriously, and I can categorically state that they consider legionnaire's disease a complete non-issue.

Yes, absolutely anything can go wrong at any time. That's why you make contingency plans, prioritized based on their likelihood. That's why, living in a place prone to hurricanes, I have a week's worth of supplies stashed in two different counties, generators (and fuel), and even portable A/C units. Along with plywood, tarps, and a decent chainsaw. And nobody is ever surprised by a hurricane!

The only way I can mitigate against a failure of an external service is to bring it in-house. For some things (eg accounting, short-term power loss) that's practical. For others, it's not.

Restarting the free accounting search

Posted Jul 31, 2017 16:02 UTC (Mon) by tialaramex (subscriber, #21167) [Link] (3 responses)

You've done continuity "planning" but have you actually had to cope ?

Legionnaire's looks at first glance much less devastating than, say, a big fire, but in both cases and many others the only way your business will actually have any real continuity is if it's inherently multi-site. That means in _practice_ not in theory there are several distinct physical locations from which all core elements of the business operate.

Most businesses are far too small for that, so in _practice_ their continuity plan for a lot of eventualities is "Go bankrupt". My favourite burrito place in London was flooded out, went bankrupt shortly after re-opening. It doesn't matter that they didn't cause the flood, or that they worked their backsides off to re-open as quickly as possible, the damage was fatal to the business.

Restarting the free accounting search

Posted Jul 31, 2017 17:49 UTC (Mon) by farnz (subscriber, #17727) [Link]

And, coming back round to the topic of cloud services, "go bankrupt" is a perfectly good continuity plan for a sufficiently unlikely eventuality (it was my previous employer's continuity plan for "what happens if our site is hit by a nuclear explosion", for example). The question then becomes whether or not a failure of any given cloud service is sufficiently unlikely that "go bankrupt' is a reasonable continuity plan.

You also have to have a reasonable assessment of the likelihood of your outsourced service failing as compared to (e.g.) your in-house service being hit by a catastrophic failure. Accounting in the cloud makes business sense if the likelihood of your cloud service failing is much smaller than the likelihood of any in-house service suffering a terminal failure.

This does, BTW, currently tend towards preferring big centralised services; if (to choose a random example) Salesforce.com goes under, your creditors are much more likely to accept that this is an exceptional event worthy of offering special terms than if your office server explodes in a shower of sparks and cannot be replaced in a reasonable time period. Not least because there's a decent chance that your creditor can't tell what you owe them until they recover from the shock, too.

As a non-business example, my continuity plan for "home hit by a hurricane" is "panic!".

Restarting the free accounting search

Posted Jul 31, 2017 17:55 UTC (Mon) by pizza (subscriber, #46) [Link] (1 responses)

> You've done continuity "planning" but have you actually had to cope ?

We came close -- Last summer, a hurricane turned a few degrees to the north at the last moment and missed us entirely. Since then, the worst has been a power outage and an internet service outage, each completely shutting us down for about half a day. (Ironically, the same things that make us more resilient to major disasters makes us more vulnerable to the more run-of-the-mill stuff..)

But your point about being multi-site is well taken; our "continuity" plans only really work because we're so geographically dispersed. Fortunately, this particular site has never had to go into full disaster _recovery_ mode.

(BTW, I am familiar with how bad Legionniare's can get, having had a rather close encounter with it in my youth)

Restarting the free accounting search

Posted Jul 31, 2017 19:56 UTC (Mon) by anselm (subscriber, #2796) [Link]

(BTW, I am familiar with how bad Legionniare's can get, having had a rather close encounter with it in my youth)

When I was a student at Edinburgh University in the late 1980s, Legionella bacteria were found in a water tank of the air-conditioning system that served the university's computing centre. This was during an early summer that by Scottish standards was quite warm, and running the computers without air conditioning was not an option. Hence the computing centre was duly shut down for at least a week – I don't quite remember the specifics – so the A/C units (all of them) could be thoroughly sanitised. While the operations staff tried to keep a rudimentary service running, this was a huge hassle as lots of people had exams and deadlines for papers or projects coming up and couldn't do the requisite work. For a company, it would probably have been even more of a problem.

So, Legionnaires' disease is not to be trifled with. I don't think anybody actually contracted it at the time, which was just as well.

Restarting the free accounting search

Posted Jul 31, 2017 14:28 UTC (Mon) by mfuzzey (subscriber, #57966) [Link]

It's not *just* about accounting service providers (though that was the original topic) but cloud services in general (non accounting services like SalesForce/Slack/Github have been mentionned).

The difference between cloud services and most of the physical services (buildings / hotels / internet connectivity etc) is that the physical services are generally commodity and it is at least *possible* (which doesn't necessarily mean easy nor cheap, especially in a rush) to find a substitute.

When the data in the cloud is *your* data (be it accounts, code or whatever), it is, by definition, unique to you and hence irreplaceable.

Now, of course, you can also loose all your data on your own servers if you don't have proper backups etc in place but that is at least your own fault.

Many people assume that the cloud means the cloud provider handles all that.
While they will (one would hope at least!) transparently handle disk crashes etc while you have a contract with them, they may not provide a way for you to get your data out (conveniently or at all) if you decide to leave them or they cease to exist.

Maybe you can, in some circumstances, sue a cloud provider who looses your data - but that won't get it back if it's really lost (provider goes broke, auctions off all the servers and the disks are erased...)

Restarting the free accounting search

Posted Jul 31, 2017 15:48 UTC (Mon) by smurf (subscriber, #17840) [Link] (9 responses)

The problem with "accounting provider in the cloud" is that, more likely than not, you are unable to either get at your data, or do anything with them if you happen to be able to download the records. If you can't generate new payroll and invoices *now*, your data dump is useless.

In contrast, if you store your database in the cloud you can trivially do a SQL dump and be back online with your own SQL server. Or move to some other AWS-compatible hoster.

In contrast², if you run the accounting system on your own hardware you can at least limp along, and you have time to look for another solution or a migration path, if/when your software provider goes belly-up.

Restarting the free accounting search

Posted Jul 31, 2017 16:09 UTC (Mon) by felixfix (subscriber, #242) [Link] (6 responses)

This was my point, however clumsily expressed. How many companies have gone out of business from cloud failures? I don't recall hearing of any. But companies go out of business from fires, floods, all sorts of natural disasters, vendor bankruptcies, landlord fiascos, embezzlers, thieves, accidents, and other causes all the time. Every once in a while I read of some company bankrupted by a former employee who destroyed their systems and put them offline long enough to destroy their business.

Worrying about cloud failures is pointless in comparison. The real world doesn't seem to have many cloud problems.

Restarting the free accounting search

Posted Jul 31, 2017 17:33 UTC (Mon) by pizza (subscriber, #46) [Link] (5 responses)

> Worrying about cloud failures is pointless in comparison. The real world doesn't seem to have many cloud problems.

The real world, however, has many _connectivity_ problems.

Restarting the free accounting search

Posted Jul 31, 2017 17:34 UTC (Mon) by felixfix (subscriber, #242) [Link] (4 responses)

And how many companies have gone bankrupt from connectivity problems?

Restarting the free accounting search

Posted Jul 31, 2017 17:39 UTC (Mon) by felixfix (subscriber, #242) [Link] (3 responses)

Let me rephrase that. How many companies have gone bankrupt from connectivity problems caused specifically by cloud failures, cloud lockout, cloud price increases, and other cloud-only failures, as opposed to electrical failures, ISP failures, cables cut by anchors or backhoes, hurricanes, floods, and other non-cloud issues?

Restarting the free accounting search

Posted Jul 31, 2017 19:01 UTC (Mon) by pizza (subscriber, #46) [Link] (1 responses)

You're splitting hairs here. What matters is that the cloud service remains available to you, not *why* it is suddenly unavailable.

It doesn't matter if the service outage is caused by incompetence by the provider, bankruptcy, hackers, puntive price increases, popular uprising, road work, solar flares, a rogue ISP in Pakistan pushing bad routes, or a gardener with a shovel. Regardless of the reason, the symptom is the same -- the service is not available, and if you need it to function as a business, you're sunk until it's available again.

One has to plan for outages.

Restarting the free accounting search

Posted Jul 31, 2017 19:12 UTC (Mon) by farnz (subscriber, #17727) [Link]

But note that "an outage of this magnitude kills the business" is a reasonable plan, if an outage of that magnitude is expected to be a rare enough event. You could, after all, plan for "aliens obliterate the continent I'm currently on", but it's such an unlikely event that there's no point planning for it.

Similarly, a small business may reasonably plan on the basis that loss of a cloud service for more than (say) 5 days kills the business, and that they'll ensure that they're close enough to up-to-date to cope with a 5 day outage.

Restarting the free accounting search

Posted Jul 31, 2017 19:46 UTC (Mon) by drag (guest, #31333) [Link]

Bad backups have caused many businesses to close their doors permanently. Some statistic I've read before says that something like 2/3rds of companies that suffer major data loss are out of business within six months.

Losing access to data or having data breach is a existential threat for almost all types of businesses. Much more serious then, say, being kicked out of a building.

If you depend on cloud providers too much and don't have a alternative way to retrieve and use your data then you are tying the fate of your business into the fate of your cloud providers. Their accidents can mean you are left dead in the water.

Even the best cloud provider has 'lost accounts'.. virtual machines go dead, customers are kicked out of their own accounts, etc. etc. Stuff happens. Their loss is your business. Your loss could be your business, as well. Your risk is much higher then theirs.

I don't have a problem with cloud services, but it really doesn't change much when it comes to dealing with backups and restores.

Restarting the free accounting search

Posted Aug 10, 2017 9:10 UTC (Thu) by oldtomas (guest, #72579) [Link] (1 responses)

Perhaps a recent hicup which illustrates this thread nicely:

https://www.theregister.co.uk/2017/08/06/cisco_meraki_dat...

(Cisco, of all things).

The other recent one I remember was also a big name (perhaps the one with the A? Not sure). Interesting -- in both cases it was (at least officially) "human error", the equivalent of the classical "rm -Rf" at the wrong time/place.

I'm waiting for the first "automatic error" -- when all those Puppets, Dockers, Lambdas and Unikernels and things are managed by some deep-learned network. That will be... weird.

Restarting the free accounting search

Posted Aug 10, 2017 18:02 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link]

You're probably thinking about Amazon S3 outage: https://aws.amazon.com/message/41926/

It didn't result in any data loss, as far as I'm aware there have been no data-loss events with S3.

Restarting the free accounting search

Posted Aug 11, 2017 15:31 UTC (Fri) by Wol (subscriber, #4433) [Link]

> Finances are not the only, or even the most, critical part of a company. That's like saying tires are the most critical part of a vehicle.

So why is it that pretty much ALL companies that suffer a serious failure in their accounting systems cease trading within the year ...

While a failure anywhere else usually only causes serious disruption for a couple of months.

Mind you, thinking about your tyre analogy, maybe tyres ARE the most critical part of the vehicle, after all, a motorway blowout can easily kill you instantly. Most other failures can be repaired.

Cheers,
Wol