What is your average download / upload speed for Insync?

[sp.peterlee]

Recently my local drive breaks down and I need all my data sync back to my computer. However during the process I find the internet connection speed through Insync is by far slower than direct connection using Google Sync or direct Download from Google Web.

Any Idea what might cause this? Is it the problem with Insync or otherwise?

Here is my internet speed.

But my actual download speed from Insync is around 15Mbps even with large files.

It doesn’t seems to be Google’s problem either, since I can obtain full speed through Google.

So what is your average Sync speed using Insync? Please share your result with me so I know it’s my computer’s problem or the best result I can get from Insync is 10-20 Mbps?

Thank you guys for helping, APPRECIATE a lot.

[jaduenas]

Our upload and download is currently limited to two data transfers at a time – one upload and one download/two downloads/two uploads – at a time to avoid saturating your bandwidth. We are releasing a workaround in the future that will allow you to change how many files you want to sync at a time.

[Hawk]

Speedtest results
Download: 15.9 MB/s
Upload: 1.36 MB/s

Insync as measured by myself (at two different times)
Download: 7.5 MB/s | 4.7 MB/s
Upload: 0.9 MB/s | 1.3 MB/s

My upload bandwidth (not so fast itself anyway) is almost saturated by Insync.
Download is not saturated, but it’s still pretty descent.

Note: I’m using MByte as unit for all results.

[sp.peterlee]

Dear Jaduenas,

I am afraid you are misunderstanding my question. I know you all have restriction on number of simultaneous upload and download. I get it !

But I am asking the download / upload speed in another word, TRANSFER RATE for the download.

I can tolerate 1 file download at a time once it’s fast like my direct download from GOOLGE. 9.8MB/Sec not 1.8MB/Sec.

**This is the problem----my total download bandwidth is not saturated by Insync. **

I see Hawk was abled to achieve good download result using Insync, why can’t I?

[Joshua_Watson]

My network speed is roughly 50 Mbps down and 6 Mbps up. My computer can access those speeds easily using Google’s Drive interface, or just about any server that can feed that much bandwidth. My Insync speeds have been as high as 45 Mbps at times, but usually it sits in the 10-20 Mbps range, even if there are no other processes using bandwidth and I’m syncing large, individual files (not a ton of read operations).

EDIT
I’m currently getting ~45 mbps on large files through insync. This was definitely not the case before deleting my database and starting over.
END EDIT

I haven’t figured it out yet. I did notice that when I finally got fed up with a corrupted database and started over with a fresh download (I have business internet or this would not have been feasible), my downloads seemed to be much faster. I wonder if changes to database structure causes a bottleneck?

Quick question - how many total files in your database? When I synced a coding folder with ~600,000 files in it, the whole program was nearly unusable, but it handles my ~150,000 files collection with only minor hiccups. I do get “out of ram” messages (I have 8 Gb) when syncing the bigger file-set (I had over 1M in there for a while, but it plain old quit working), so I wonder if it’s attempting to cache and then page a bunch of stuff and subsequently getting stuck on the latency of my storage on HDD computers? That would mean the bottleneck is not bandwidth, but read speed or I/O operations. I don’t get the “out of ram” or slowness issues on my SSD computer, which lends credence to this idea.

[sp.peterlee]

I am doing a fresh sync, and due to slow sync process, I haven’t finished a fresh sync as yet. I am running 8 enterprise HDD on a RAID 6 so I have pretty good I/O and read/write speed for the logical volume. I use RAID solely for storage so no other tasks are being performed during syncing process.

I have 1.07 TB of data and total of 319211 files syncing. I have 16 GB of RAM and no such “out of ram” messages ever pop out.

Thank you for your input. I will continue to observe Insync’s activities and sycing speed when I tried out large files.

By the way, Great Job Insync for your next futuristic promise, I am sure reading the post will solve my bandwidth issue.

[Joshua_Watson]

Firstly I’m pretty jealous of that lovely set-up. =) Secondly, we both agree that there’s no way that I/O is the issue for you.

RAM is probably not the issue with 16 Gb. My conclusion is that the problem is one of the following:

• Throttling to avoid network saturation is “guessing” an incorrect “maximum saturation” value, which causes the throttling to be too strong. This one could be fixed by allowed users to push a button in settings to temporarily remove all throttling for a “time sensitive quick sync.”

• The scanning algorithm (or accompanying interaction with the database files) is overflowing or hanging because of the large amount of files. The --skip-local-scan command line argument might temporarily fix that - I haven’t tried yet.

• Something wonky was going on with the database, possibly due to in-place updates. This doesn’t seem to be your problem - did you delete the database before re-syncing?

I really appreciate good troubleshooting - thanks for your contributions @sp.peterlee . I’d love to not worry about this in the future.

[benmeroff]

Any update on this? I’m still only getting around 20-30 MBps upload speed on large files when I typically get around 200-300 MBps upload speed on our network for everything else; is there a Google Drive upload speed cap in play here, or is it something due to Insync’s API?

This also has no difference when uploading one or two files simultaneously, my upload speed per file will always stay right around that 30 MBps at maximum.

[simong]

I’m never had a problem with transfer speed from the client

I’ve got 1Gbit connection (http://www.speedtest.net/result/6250082923)

And one single transfer (a 2GB in the example below) usually transfers at somewhere between 400-800Mbit/s.

But the max 2 threads is the biggest problem, especially when there is a lot a small files in que. A 100KB file compared to a 50MB file takes the same total time to upload since the most of the time is internal insync processing time, checking file, updating databases, get the next file in the que etc etc.