08.Oct.2011 at 16:23 by Rob Allen via api
Interesting talk about the challenges faced when handling lots of data.
PHP and Continuous Data Processing
Michael Peacock (08.Oct.2011 at 11:45)
Talk at PHP North West 2011 (English - UK)
At Smith Electric Vehicles we process on average 1 billion pieces of information in our PHP based telemetry application. This is the largest vehicle telematics project in the world with the exception of Formula 1. With this information we need to be able to display live vehicle metrics to customers, plot vehicles on maps, trigger alerts when diagnostic flags are raised, process the data to generate reports, and export large volumes of this data for various stake holders, all without losing a single piece of data.
This talk will discuss some of the technologies and concepts we employed to ensure the system is reliable, fast, and can be maintained without losing data. In particular how we keep the web application running quickly with all of the data behind, how we sustain the huge number of inserts into our database, how we can analyse the data and how we export the data needed for monthly reports.
Quicklink: https://joind.in/3808
Track(s): Unconference
Who are you?
By clicking this button you are declaring that you are the speaker responsible for it and a claim request will be sent to the administrator of the event.
If the claim is approved you will be able to edit the information for this talk.
Are you sure?
Comments
09.Oct.2011 at 20:25 by Jenny Wong
A well timed presentation. it was interesting listening to the problems and the solutions that they are using and the reasons behind it. No doubt could be extended but it did give a taster of things to come.
It may be a preference of mine, but would have loved to see a little more passion + personality flaring from Michael.
09.Oct.2011 at 22:27 by Remon van de Kamp
Really interesting to see what problems Micheal is facing in his job, and very impressive with how's dealing with them. He seems to have things very much under control. Presentation was good, but it was pretty clear he was a bit nervous. That will probably go away as he gives more presentations on the subject (which he should, because it's very interesting).
10.Oct.2011 at 16:43 by Alex Butter
Very interesting subject matter, presentation was a little too short though. I think Michael could easily fill an hour on this subject, although that might get quite hard-core as far as the content is concerned (i.e. getting pretty deep into the nitty gritty coding instead of the somewhat more general overview presented here).
I've certainly picked up some ideas which I'll be using in a coming project where we're also processing large volumes of data continuously (although not on the same scale as Michael's team is).
Excellent!
11.Oct.2011 at 12:37 by Jamie Hurst
Brilliant overview to the topic of data processing, especially interesting to see how some of the more intricate parts of the solutions at Smith's are implemented, and a nice look behind the scenes too.
I agree that this could've been a full presentation though, an hour on this would've been fantastic! :)
08.Oct.2011 at 11:04 by David Goodwin via api
More time would have been nice; to have gone into more depth in places .
Nice range of Issues and solutions. Thank you.