[SOLVED] 2 x 10 gig port server.

[gyke69]

Hello!

i would like to know is there possible way to connect 2 x 10 gig port server together so i have 20 gig port connect? if so please clarify how am i able to achieve that.

 

[kanewolf]

Hello!

i would like to know is there possible way to connect 2 x 10 gig port server together so i have 20 gig port connect? if so please clarify how am i able to achieve that.

You can use link aggregation (LAG) if you have a switch that supports it. You could have a total of 20gbit, but individual connections may still be limited to 10gbit depending on the host LAG implementation.

 

[gyke69]

You can use link aggregation (LAG) if you have a switch that supports it. You could have a total of 20gbit, but individual connections may still be limited to 10gbit depending on the host LAG implementation.

i am literally talking about VPS server not a dedicated server at home or office.

 

[bill001g]

This is one of those talk to the provider. It all depends on how they simulate the bonding and what actual hardware exists between the virtual machines or if that too is simulated. In many cases the network is not the bottleneck when you have even 1 10gbit port. Many times it is the disk system or the file sizes and structure that limit the transfer rates.

 

[kanewolf]

i am literally talking about VPS server not a dedicated server at home or office.

How would I have known that based on your original post?

 

[palladin9479]

i am literally talking about VPS server not a dedicated server at home or office.

So is this physical or virtual, huge difference between them?

Now I'm assuming so far that it's a virtual server you have on someone's cloud and that server has been configured with two 10 Gbe virtual interfaces? If so then there are two answers, the first is the same as with a physical server, the other is will also work but is kinda janky.

First is that if your cloud provider supports LACP (Link Aggregation Control Protocol) on it's virtual network infrastructure, AWS calls it LAG (Link Aggregation Groups) but it seems more geared towards site to site communication so may not be available for host ports.

Second method is janky and ghetto, but works. Ensure your service is listening on both interfaces, then configure a DNS entry with both interfaces IP address's in it. This will create a form of load balancing known as "DNS round robin" where each client connecting will get a different IP and connect to that interface instead of everyone going to a single one.

The "correct" answer is not have multiple 10Gbe on a single virtual server and instead setup multiple instances with some sort of load balancing / clustering. Having two servers (or more) running the application with a load balanced placed in front of them is the industry wide accepted practice. Can replace servers with container node for those running containers. This way not only do you have higher bandwidth, but also service redundancy and possibly zero downtown maintenance.

 

[gyke69]

How would I have known that based on your original post?

Now you do. ))

 

[gyke69]

So is this physical or virtual, huge difference between them?

Now I'm assuming so far that it's a virtual server you have on someone's cloud and that server has been configured with two 10 Gbe virtual interfaces? If so then there are two answers, the first is the same as with a physical server, the other is will also work but is kinda janky.

First is that if your cloud provider supports LACP (Link Aggregation Control Protocol) on it's virtual network infrastructure, AWS calls it LAG (Link Aggregation Groups) but it seems more geared towards site to site communication so may not be available for host ports.

Second method is janky and ghetto, but works. Ensure your service is listening on both interfaces, then configure a DNS entry with both interfaces IP address's in it. This will create a form of load balancing known as "DNS round robin" where each client connecting will get a different IP and connect to that interface instead of everyone going to a single one.

The "correct" answer is not have multiple 10Gbe on a single virtual server and instead setup multiple instances with some sort of load balancing / clustering. Having two servers (or more) running the application with a load balanced placed in front of them is the industry wide accepted practice. Can replace servers with container node for those running containers. This way not only do you have higher bandwidth, but also service redundancy and possibly zero downtown maintenance.

I was looking for this answers.
Thank you.