SaaS data loss: The problem you didn’t know you had

As Australia's heat-wave returns, I’ll be in the Blue Mountains in NSW, surrounded by eight hectares – around 20 acres – of mostly uncleared bush, contemplating both the benefits and shortcomings of modern emergency services communications.

Here with two smartphones and a laptop with a mobile broadband connection, there are plenty of ways to watch what’s going on. However, they all fall short in some fashion or other.

The laptop is useful for checking new fire updates posted at (since I’m in NSW) the Rural Fire Service’s incident map. However, it’s rash to rely solely on that site. If an emergency breaks out, it can struggle (Victoria’s Country Fire Authority site crashed under a heavy load on January 3).

State on fire: the NSW RFS Incidents page on 10 January, 2013

There are alternatives. Google offers a mirror here, and there are individuals’ efforts like this (which is mapping fire advisories from all states).

The problem is that people don’t sit at their computers all the time. Sure, the sites can be browsed from a mobile, but that’s inconvenient and illegal if you’re driving a car.

Which brings me to the second modern emergency advisory: the mobile phone. SMS alerts are sent to people living in the path of a catastrophic fire.

The system works, but it has its own limitations. Coverage outside Australia’s big cities can be spotty. If you live away from mobile coverage, the SMS alerts might not reach you. If you’re a tourist passing through a danger zone, you probably won’t even be on the alert list.

This is known: Victoria’s acting premier has made a public statement that people should not rely solely on SMS alerts. He’s been backed up by that state’s Fire Services Commissioner.

Another issue that affects both Websites and SMS alerts is the timeliness of the information. We don’t have live, 24-7 remote sensing. Satellites like MODIS can tell us lots about the progress of a fire on a daily basis, but live information comes from people on the ground.

These are the same people whose main preoccupation is fighting the fire – which is why I was interested in the idea of using UAVs to help map fires in real time.

Even drone-gathered information would, however, have to get from the ground to the public – which will mean someone has to check it for accuracy before release. Since we don’t have the drones yet, we’re stuck with today’s approaches.

As well as the people actually standing in front of the fire-front (raise a hat to all bushfire volunteers!), there are aerial spotters in planes and choppers, fixed watchers in towers, and (where they’ve been installed) remote cameras.

Information from hundreds of fires has to be collated at headquarters, checked (because nobody wants to spark a panic or, worse, send evacuees in the wrong direction), and published, with appropriate decision-making by senior personnel and contact with police and other authorities.

On hot, windy, days like those we're currently experiencing in Australia, any of the hundreds of fires still burning could take off suddenly. It’s not hyperbolic to state that a fire can escalate to a catastrophe in minutes rather than hours. A strong wind can also send glowing embers many kilometers from the fire-front.

All of this can happen far quicker than the supposedly-instant online world can react.

What about social media? Shouldn’t I also turn to Twitter to look out for dangers in my vicinity?

Yes and no.

I know who to look for – people who live in the right place to see the smoke or the flames.

I don’t follow the Rural Fire Service Tweets because they’re most often links to the incident page that I already watch.

The rest of Twitter is as much use as two hundred kilometers away: people Tweeting their well-wishes are nice, but not informative; people re-Tweeting obsolete warnings from four hours ago merely noise, as are those who clog the feed with links to news stories.

The CSIRO has told the Australian Financial Review that 12.5 million tweets relating to the bushfires have been sent this week. That's great for someone able to cope with the "big data" analysis, but it renders the medium useless to the individual.

Mainstream news publishers are especially guilty of feed-clogging. If I wanted to rely on Twitter as my bushfire information source, I’d have to find some way of muting media feeds that are either image-feeds (they don’t actually say “Send us your pics of the catastrophic fires now engulfing your community!”, but they may as well) or news stories that are running half-an hour behind ground truth.

All of which means my most important technical aid on an extreme fire danger day is also the oldest one to hand: an AM radio tuned to the Australian Broadcasting Corporation’s local service, present both in the house and in the car, and able to break into its programming in real time when there’s an emergency.

I really can’t see that changing for years. ®

Regcast training : Hyper-V 3.0, VM high availability and disaster recovery