Repeat after me : Spark is my friend.

[mac]

Hello,
for those who are, like me, paranoiac on the internet, just go to the log file of your firewall and check connexions launched by spark.
Hundred and hundred of connexions (that were blocked) pointing to ip adresses 52.85.180.*** wich belong to Amazon ; the same to ip adress 52.185.218.*** wich belong to microsoft.

No more comment.

[jvi]

Yes, it could be more less...


Verzonden vanaf mijn iPhone met Tapatalk

[ogs]

That is a very strong argument for NOT using spark. Unnecessary connection attempts to the internet is not good. Devialet has some explaining to do. I am not yet using Phantoms, but if Roon can play directly to Phantom or Dialog at some point I'm interested. Then I would not have to use Spark..

[andrewhollo]

Hello,
for those who are, like me, paranoiac on the internet, just go to the log file of your firewall and check connexions launched by spark.
Hundred and hundred of connexions (that were blocked) pointing to ip adresses 52.85.180.*** wich belong to Amazon ; the same to ip adress 52.185.218.*** wich belong to microsoft.

No more comment.

What is the significance of this? Why does Spark do this? Why is this a problem?

Andrew


Sent from my iPhone using Tapatalk

[ogs]

It is a problem because we do not know what kind of information Spark is attempting to deliver to Amazon and Microsoft. If you have not configured anything to connect to those two, why would Spark do this? Why must we study firewall logs to find out? Amazon and Microsoft has nothing relevant to offer in local steaming to Phantoms in my view.

[raph972]

@mac
Are you using any service like Deezer, Qobuz or Tidal ?
My guess is that when you're browsing any of those services, or retrieving cover art and such, you have to communicate with those services cloud backends which are very likely to be hosted on Amazon Web Services, Microsoft Azure or Google Cloud as this is what all the cool kids are doing now.

In that case, it is completely normal for the app to communicate with those servers to provide you with what YOU requested from the app, right ?

[Zappydev]

@mac
Are you using any service like Deezer, Qobuz or Tidal ?
My guess is that when you're browsing any of those services, or retrieving cover art and such, you have to communicate with those services cloud backends which are very likely to be hosted on Amazon Web Services, Microsoft Azure or Google Cloud as this is what all the cool kids are doing now.

In that case, it is completely normal for the app to communicate with those servers to provide you with what YOU requested from the app, right ?

Roon will also communicate extensively in the background; not only to download metadata, but also to upload information about the usage of that metadata which is used to allow the metadata providers to charge Roon the appropriate fee for that usage. I believe that goes via an Amazon server. See more on the Roon forum.

[ogs]

@mac
Are you using any service like Deezer, Qobuz or Tidal ?
My guess is that when you're browsing any of those services, or retrieving cover art and such, you have to communicate with those services cloud backends which are very likely to be hosted on Amazon Web Services, Microsoft Azure or Google Cloud as this is what all the cool kids are doing now.

In that case, it is completely normal for the app to communicate with those servers to provide you with what YOU requested from the app, right ?

Roon will also communicate extensively in the background; not only to download metadata, but also to upload information about the usage of that metadata which is used to allow the metadata providers to charge Roon the appropriate fee for that usage. I believe that goes via an Amazon server. See more on the Roon forum.

All correct. We know Roon is talking a lot to services on the internet for metadata and usage. Redundancy across providers is probably a good thing, but Spark does not depend on this, does it? Not to my knowledge anyway.

[booshtukka]

Just as raph972 says: they are almost certainly hitting Amazon and Microsoft for the AWS and Azure services respectively, which act as a kind of CDN (content delivery network).

This is basically like a huge hard drive on the internet where you can store stuff (in the cloud), and then deliver it quicker to the end user because you can choose to locate that hard drive physically closer to the user.

There's nothing sinister about those IP addresses. I don't know what Spark is making those connections for, but those are not shocking or unexpected endpoints.

[ogs]

Just as raph972 says: they are almost certainly hitting Amazon and Microsoft for the AWS and Azure services respectively, which act as a kind of CDN (content delivery network).

This is basically like a huge hard drive on the internet where you can store stuff (in the cloud), and then deliver it quicker to the end user because you can choose to locate that hard drive physically closer to the user.

There's nothing sinister about those IP addresses. I don't know what Spark is making those connections for, but those are not shocking or unexpected endpoints.

I would not use shocking, but the OP reports that his firewall blocks these connection attempts. Spark seems to function properly even if this is blocked. In that case the connection attempts should not be there. When @mac enables one or more of the services that require a connection to AWS or Azure it would not work as long as his firewall is blocking these connections. He'd then have to open up his firewall for the specific traffic to make it work.
Most people only have a simple NAT between internet and their home network. Then "everything" will work, including malware and bot-nets. A next gen firewall to protect your home network is a good thing in my view.