As I was saying…

• Posted On: Saturday, September 18th, 2010
• Post Categories: Blog

It recently occurred to me that one of the fringe benefits of WAN acceleration often goes unmentioned: resuming broken uploads.  Recently we had a customer ask if we could help with a new service they were developing – a sort of in-house Youtube. One of the challenges they have is that many of the sites participating in this service are served by 512Kb so uploads are going to take a long time and it’s likely some of them will fail or time out.  Object-based caches don’t handle this kind of failure.  Obviously there are tools, including ftp, which do recover broken uploads but if you’ve already got Replify Accelerator in place, you don’t need any of them.  All of the previously sent data will be sitting as cached blocks in the destination virtual appliance and so a second attempt to send the file will rattle through that data sending short references to the blocks and in only a few seconds you can have recovered tens of minutes of the previous upload.  Even better, if you cancelled that upload to make one final change to that presentation or whatever (how many times have I done that!)  you’ll still get the benefit despite this now being a new version – because most of it will remain unchanged.  In case you don’t believe me, I’ve recorded a short video to prove my point. I’ve chosen to use SharePoint as the target system but it could be a fileshare, or anything else. You can watch the demo here: