Faster uploads and downloads

[burnandreturn]

One of the biggest drawbacks to using Fusion 360 is the slow uploads and downloads of anything to do with Fusion360.

Evern this website loads and updates at a snails pace.  although I don't have a blazingly fast internet connection it works fine with every other online expirience.

A friend of mine asked me to check out OnShape for him.  My experience so far has been very good.  No crashes and everything is cloud based and it runs fine and is fast.   So if you insist on living in the clouds then it can be done and be not a terrible expirience if it is fast enough.

Trying to import  models from other software into Fusion is taking way too long.   I asked a question recently on the forum on how to cancel uploads that I have tried just because it takes way way too long and is not reliable.   Fusion often crashes or just stagnates with uploads.

I find myself closing out of the program because it is so slow to transfer to and from the cloud.  I don't know if it is just the sheer numbers of users that is bogging down the servers but something is not right.

Maybe you should ask FaceBook how they handle the server loads!

[HughesTooling]

What file types are you trying to upload. You can use New Design From File to load stp, igs, sat, smt and Fusion's f3d files without uploading to the cloud. Also part of the upload time is for translation, if you just upload a file in Chrome with the Dashboard with out translation it's quite fast, so the translation part needs the speedup.

Mark

[roambotics_scott]

+10

I always assumed that this was a weird issue with how F360 interacts with S3, but if the site and the translation are running on EC2 instances, perhaps it's just the instance types AC is using ?

If you're using any of the t2 (even t2.large), there can be a novelty amount of latency. 

I'm almost certain that they deliberately cripple performance on the cheaper instances (read: everything below ~xlarge or r3.xlarge ) or / at the very least, they're running in VMs and not in dedicated hardware, so if they're not actively in use, there can be a significant delay before they fully wake up.

Just swapping out the instances could be a quick fix.

Translation might be an interesting place to use lambda (likely with better performance, lower latency, and for significantly less money if the instances are currently underutilized)

[colin.smith]

This item has been archived due to a lack of votes. If you feel this item is important feel free to create a new idea station item.