Large number of Laptops to be backedup

Hi

We can give you some advice, but looks like you are asking for a use case, and some sort of support.
Better to open a case for this, and chat with a TAM.

I can give you some advice, 
more than one proxy, in my opinion, two and the VBR, for haqndling the tasks.
Divide the coputer in ingest groups, 30 - 50 laptops each.
Make sure the cache is enable on the laptops, for performance and wifi issues.

cheers.

Basically when you use a Veeam Agent, the agent does the workload and actually moves data to the repository. I’m pretty sure, that WiFi is your bottleneck here. Maybe your WiFi itself is at-limit when you run multiple Agent jobs at the same time (many laptops backing up at the same time).

The proxy is the Agent in this case, the VBR server is the component coordinating the config.

I would go with a support case as well to check if there are any issues. Otherwise, I advise to reach out to your Veeam partner (VASP) or Veeam sales rep to get some architectural reviews and recommendations.

There’s a lot to unpack here that is unsaid.

Are they backing up in parallel? Or have you split them into batches to run across the day?

How many APs are these devices split against? Remember Wi-Fi is unicast so your devices could be contending for airtime before the traffic even reaches Veeam.

I assume this proxy is your gateway to S3 storage you mean? How is it sized? Are all devices on the same network as the proxy or are these machines also traversing either some site to site link such as an MPLS or VPN? Are there remote endpoints that are VPN connecting back into your network?

How much bandwidth do you have available between your site(s) and the S3 storage? Is it a Local S3 storage or something cloud based such as VDC Vault or Wasabi?

A lot of questions as you might gather to help find out where to start optimising

I agree with both takes here ​@Rendym . Reach out to a Veeam TAM or Support to get Job design assistance. But no doubt your wifi is more than likely the bottleneck. Wifi has more limited bandwidth than hard-wired connections.

In the Backup Job history (double-click the Job), at the bottom of the task list (Action column) in the right window, what is shown as the bottleneck? There are 2 tasks to take note of at the end of the Job → “Load: Source: x% > Proxy: x% > Network: x% > Target: x%”  ; and “Primary Bottleneck: <one of the ‘Load’ options will be listed>”. What are the given percentages and what is Veeam stating is the bottleneck. From these 2 items is where you start your performance troubleshooting.

Best.

Below is an example what I’m suggesting you look at:

The above is an example of a Replication Job I have. I use “hotadd” as my Backup Method, and this of course goes across a long-range LAN with fiber between, but even so...network is still my bottleneck because of the distance the data traverses. But, I don't really have performance issues. The Job runs pretty fast.

Hope that helps.

400 laptops backed up over Wifi?  I have 800-1000 Mbps WAPs (Wireless Access Points).  125 MB/s @ 1000 Mbps is the maximum speed.  Most laptops get 400-460 Mbps.   

1.) Do you want to backup the entire laptop for recovery?  Or just documents, pictures, and work files? 

2.) How many WAPs to you have?  Are each on gigabit connections?   

Hi ​@Rendym 

From a Desgin perspective, will understand backing up all the laptops in one backup and one protection group is great, in practice it’s just don’t really work. 

Taking a step back a looking at what your trying to accomplish from a whole ecosystem perspective there are 

Some of the biggest Desgin issues you’re going to encounter is concurrency and airtime fairness across the access point, the bottleneck with at switching levels and all the network are going to be tagged at best effort means that a big queue of packets all trying to reach the same s3 destination at same time. 
 

Your internet provider in your offloading to External provider s3 provides will take the network pack from fifo format, which Potentially means you mean you will re transmissions as network packets arrived in out of sequence, this all case delays in the overall time it takes to complete the backup
 

depending on the s3 provider you may an encounter API throttling is it tries to process your traffic through its API model into S3 bucket. My assumption is that you’re using assumption for all laptop backups. 

My recommendation if at all possible is to breaker protection group into smaller amounts and have different  backup jobs for each per Protection group

this then allow you scheduled at different times of day in an effort alleviating the congestion issues that you are going to inevitably face.