Hi Sue,

I couldn’t see a download to a file that you have uploaded on that page though. Is there meant to be one?

I was wondering if it would be possible to see an actual link to a file download. I wouldn’t actually click on it to download and waste bandwidth though, rather just curious as to how it all works.

Best wishes,

Mark

Hi mark,

Replying to your earlier post…. it was never my intention to have an entire EE site residing on S3, only select media. Images and any small files would all still reside locally. In my case I’m intending to put video content up there.

So certainly any expansion we work on would be extending the way you work already rather than turning it on it’s head.

As for how easy it would be to do what I’m intending with EE framework, I really can’t say. I know LinesDrawn already had something working, and at some point he was going to share that code so we could build on it. So I’m sure getting S3 working will not be major.

Our intention is very much to make S3 more a feature of a large expansion of the EE upload feature. I’m just not sure to what extent we can plug into this, as I’m not the developer, I will simply contract the job out.

Here’s a file you can download… and feel free to download it

The thing is, the way I have it setup you don’t know it’s Amazon S3. I have it hooked into a subdomain. I can explain the technical side of setting it up this way if you want. But let’s just say it’s shockingly simple.

Goto… http://files.moja.com.au/robs.mp3

This one is hosted in the US DC, I don’t have anything in Europe at the moment, can put one up if you like though.

You don’t have to setup a subdomain like I have above, you can link directly to to an Amazon URL if you prefer.

If you use a Mac, check out FTP clients Transmit or Interarcy. Both support S3 very well, in particular Trasmit.

Amazon supply an awful lot of sample code to get your own app up and running, and if you look at their site, there are a growing number of applications appearing that exploit S3 in different ways.

Hiya,

Yes I thought that providing a whole site would be a bit much. I too am looking at just having files that can be downloaded stored on Amazon and the main EE / html of the site stored on a normal web-server. That way I offload all the storage / bandwidth required for file downloads to Amazon and keep the site running nice and lean.

Any more information anyone has on doing this type of setup would be massively appreciated.

Best wishes,

Mark

Mark

PDF Available for EEDocs is a link to a file that is stored on Amazon.

It’s no different than linking to anything else on another server…

Hi Sue,

Thanks for that. Very kind of you to show me the link. I will look into the usage of this right away!

I feel maybe another plugin or extension coming on in the near future!

Best wishes,

Mark

Yes well the plugin is only going to make it that much easier. I’m doing it now, uploading to S3 using transmit and linking to the image normally. Works fine for me, my motivation for the plugin is to open all this upto my less tech savy clients.

Another easy way to connect to an Amazon S3 account… http://s3browse.com. Once you have signed up for your S3 account, you can upload files via this website using a web based client works very well actually.

We actually have some of our sites setup now so the larger files are all on S3, the server just holds the EE code and smaller files. With the EE database hosted on a separate MySQL cluster. It’s an amazingly reliable and cost effective setup that brings a lot of redundancy into the mix behind the scenes, for not much more than a regular decent hosting plan.

neekster - 01 February 2008 10:17 AM

Yes well the plugin is only going to make it that much easier. I’m doing it now, uploading to S3 using transmit and linking to the image normally. Works fine for me, my motivation for the plugin is to open all this upto my less tech savy clients.

Another easy way to connect to an Amazon S3 account… http://s3browse.com. Once you have signed up for your S3 account, you can upload files via this website using a web based client works very well actually.

We actually have some of our sites setup now so the larger files are all on S3, the server just holds the EE code and smaller files. With the EE database hosted on a separate MySQL cluster. It’s an amazingly reliable and cost effective setup that brings a lot of redundancy into the mix behind the scenes, for not much more than a regular decent hosting plan.

Hi Neekster,

Where in Australia are you? I am based in Cairns and from Slough originally. Can’t be happier!

Is there any word on the plugin for S3 services? I am looking to use a similar system for a national archive website. I also need to know more about downtime plans and backup as I’ve never used them for website content before.

Is it easy to extract the URL for the resource and is there also a way to prevent leeching?

Thanks for the help!

Hi Stephen,

It’s funny, I thought you knew who I was, it’s Nick from Moja. So knowing that you would probably know I’m a Perth boy! That’s great your loving Cairns, once of the few areas I have yet to visit.

That being said, why I assumed you knew Neekster = Nick I’m not sure, guess my brain works in mysterious ways.

Back to the topic…

I’m still very keen to get this S3 plugin out the door, although it’s probably slipped from the top of my list due to the lack of interest.

We could do a pure S3 plugin tomorrow, but really the end user cares less where their files are stored, which is why I think it needs to be a complete replacement of the upload functionality so as to handle everything in multiple storage options. As in the case of videos for example, you may want the Video stored on S3 and/or EC2, and store the thumbnail on faster, local storage. So a simple plugin to S3/EC2 may not work very well in this situation.

So on the basis of that, I really would like to replace all the upload functionality. Although from what my developers tell me this would be more a modification of the EE code rather than a plugin.

So really need more feedback from the wider community on what everyone else would want, otherwise the public will only get what I want!

As for some of your other questions.

Extracting the URL is not a problem, you really dictate what it will be. The files can live on the same domain as the website by using a sumdomain. In our case we set S3 files upto live on files.yourdomain.com, so nobody will ever know they are coming from Amazon.

Hotlinking is not something I have looked at, I can’t see any way that amazon natively supports this, although I’m sure some custom code could be put together to achieve this.

I assume by downtime plans you mean achieving high uptime for a particular site. It’s a bit off topic for this post, so perhaps make another post detailing your needs a bit more. I do have some experience in this area, but it is something that’s much harder to achieve in Australia. Firstly you just won’t achieve 100% uptime, but you can get very close. There really is very few if any host’s in Australia that cater for this. There is a certain Perth host who I shall not name that trumpets their clustered hosting service. But when I talked to them about 18 months ago now, they only had a single carrier servicing their DC. What’s the point in having clustered hosting if everything is going off the air when that single carrier goes down?

My point? You really need to look at every conceivable point of failure to get even close to 99.99% uptime.

I could keep going on this topic all day, which is why it’s better in a separate thread!

Hi neekster,

any news on the S3 plugin/module? Personally, I’d just like to be able to upload files to S3 and get a URL back that I could then save in a custom field. For example on one site I’m working on, we need to upload several files every week (PDF, about 50 of them) and add them to individual entries. And trust me, uploading them with Transmit, copy-pasting the URL 50 times is exhausting to say the least. It would be a giant time saver to be able to complete the process in less steps.

Hi Davus,

No good news I’m afraid. The project I wanted it for ended up not really having the budget to make the development worthwhile. We were interested in building the module to give a bit back to the community, but there just didn’t seem to be enough people interested.

Just a suggestion though, why not hard code the URL into your templates or as a global variable then just use a custom field to pass only the filename of the file… still copying and pasting but at least you would only have to worry about the file rather than the full URL.

Hi neekster,

ok, that’s understandable. Would’ve been fun though.
And thanks for the suggestion, I was already thinking about a similar solution. Actually, I wanted to name the files according to weblog and title, so I could get rid of the copying alltogether. We’ll see how that turns out.

I’m still keen to see this wonderful S3 module, if it gets made…

Here’s my question to the group:  What would be your thoughts about having ALL images on S3.  Would that enable a website to be more scalable (and withstand traffic spikes / digg effect) or would it cause other problems?