Linux Code for “Device Memory TCP” – Network to/from Accelerator RAM

https://www.phoronix.com/news/Linux-Device-Memory-TCP

by ksec • 3 years ago

80 45 3 years ago

arjvik• 3 years ago

Somehow I think the overhead of TCP will be a lot more significant than the overhead of a copy from cache to memory.

Wouldn't such a system make more sense for sending raw memory directly over the wire in some form of custom protocol? Even UDP would probably be able to approximate that.

lionkor• 3 years ago

TCP lends itself well, I assume, because it:

- is supported everywhere

- allows massive reads and writes (whereas UDP often has quite small limits before packets are dropped)

- is reliable (you wont lose data, your data won't corrupt, etc)

So you could use this inside an existing datacenter, with existing hardware, and existing software to debug and monitor it.

If you send GB of memory, the TCP overhead is minimal.

klabb3• 3 years ago

Tcp is also extremely battletested and performs well not just theoretically but in the face of other traffic (congestion) and in many different environments.

Beating TCP reliability and performance is an extremely ambitious endeavor in practice. Add wide interop to that list and it becomes next to impossible.

toast02• 3 years ago

Have a look at Netflix's 800Gbps CDN node presentation. If you want to hit high transfer speeds, your CPU may not have the bandwidth to read from your device and write it to the NIC. DMA between devices (assuming that's what is happening here), saves memory bandwidth for other stuff.

rewmie8• 3 years ago

Have a look at Netflix's 800Gbps CDN node presentation. If you want to hit high transfer speeds, your CPU may not have the bandwidth to read from your device and write it to the NIC.

Would it be feasible to horizontal scale those nodes instead?

londons_explore• 3 years ago

I assume the actual TCP packetization is done by the network hardware in the typical case. Then the overhead of TCP is hidden as long as the network hardware can operate at full line rate.

benlwalker3• 3 years ago

At 200+ Gbps, the copy from LLC where the packet landed to the userspace buffer dominates the performance profiles on most systems. The TCP processing isn't bad and the expensive parts can often be offloaded. I'd contend that this data copy on the TCP recv path is one of the most important performance issues to solve for the entire industry right now. DDIO let us kick the can way down the road, but it seems like the network speeds have outpaced the CPU speeds so much that it can't save us for much longer.

eep_social• 3 years ago

I think performance limitations are addressed directly by another comment but also consider that UDP might not be available.

jand• 3 years ago

Devmem TCP sounds a lot like direct memory access. Am i mistaken if i think of it as a security nightmare? Do you have by chance any links to security considerations?

abwizz• 3 years ago

i think it's intended for use in private networks (not the internet) like san's

egberts12• 3 years ago

I created a user-mapped array TCP-devmem variant for 2.14 in which packets can be written from userland in zerobuf-style.

Security is/was pretty good once an LSM permits the connection creation.

rcxdude2• 3 years ago

DMA is only a problem if transfers can be initiated by an untrusted party. This patch looks to basically just be making it so that transfers which would normally go device->RAM->device with no actual processing of the data transferred instead just go device->device instead (but still initiated by the OS, not the device or the network!). This doesn't affect access control at all, as the trip to RAM doesn't impact that (though it does mean it only works if the machine is not looking at or touching the data at all, which probably means it's only useful in situations where there's no encryption).

buildbot2• 3 years ago

How is this better than RDMA that is already well supported?

At least with mellanox stuff the last time I touched it, there were I think four levels of reliability, one of which matched TCP.

fulafel• 3 years ago

This sounds like it will work with existing TCP applications.

yokaze• 3 years ago

It's different. Rdma offloads the TCP transfer of host memory to the network device. This is about transferring device memory directly, avoiding going though the host.

Had to read through the lklm message, at it wasn't clear from the article.

_zoltan_• 3 years ago

RDMA does not offload TCP. Furthermore it's not limited to between NIC memories - for example GPU memory to remote GPU memory works as well.

justinclift• 3 years ago

From what I remember of Infiniband, the network cards are able to do direct PCIe transfers to/from compatible devices attached to the host (eg nvidia quadro gpus).

Specifically to avoid bogging down the host.

KaiserPro2• 3 years ago

This is my first thought, it sounds like inifiniband with with none of the guarantees.

I suppose it means that you can bust out of the local network, and take in "foreign" data. But if you're doing that, you've got bigger problems with data affinity.

ta8645• 3 years ago

Perhaps it's just avoiding the need to encapsulate an InfiniBand transport packet? I'm not sure how much overhead that introduces in RDMA, but DMTCP may just be an effort to divorce the idea from its original IB underpinnings.

benlwalker3• 3 years ago

RDMA is great and similar, but behaves very differently from TCP in the face of network congestion and longer distance traversals. This is essentially trying to get the best parts of TCP and the best parts of RDMA combined.

buildbot2• 3 years ago

Wouldn’t that be RoCE? Or iWarp?

armitron• 3 years ago

The response [0] from a Linux kernel maintainer means this is dead on arrival.

[0] https://lore.kernel.org/dri-devel/8719bf65-6906-0426-1ce4-b0...

rincebrain• 3 years ago

(I work for Google, not on anything related, opinions my own, etc)

I would suspect it's not going to be dead unless the maintainers think the fundamental premise, not the architecture, is broken - if it's been brought up on a real piece of hardware, it's almost certainly because it was solving a real problem being experienced, and even if it involves reworking it completely, getting it upstreamed is much better than maintaining an enormous local delta.

rewmie8• 3 years ago

To sum things up:

I strongly suggest that you read up on the basic of this before proposing any solution.

Ouch.

pantalaimon• 3 years ago

Eh that sounds more like the API was abused/used wrong - not that the whole idea is fundamentally flawed.

That’s just what code review is there to spot, not something that can’t be fixed.

jeffbee3• 3 years ago

My experience watching google try to contribute to upstream kernel is their proposals are almost always based on things that have already been in production at google for years. So when I see an upstream gatekeeper accusing someone of not understanding what they are proposing, I am skeptical. The thing that is “dead on arrival” here is probably the ability of the general public to enjoy something that google already uses.

rewmie8• 3 years ago

So when I see an upstream gatekeeper accusing someone of not understanding what they are proposing, I am skeptical.

Your comment reads like a textbook example of an appeal to authority.

Google engineers don't walk over water, and what Google accepts in their internal networks considering on Google's tightly locked internal security infrastructure might not map to real world security environments.

It also boggles the mind how you somehow presume a Linux kernel maintainer who repeatedly pointed out the failures of the proposed patch does not know what he is talking about. Unbelievable.

The thing that is “dead on arrival” here is probably the ability of the general public to enjoy something that google already uses.

I'd wager that what the general public enjoys is working code with an excellent security track record thanks to the outstanding work by Linux kernel maintainers, who repeatedly stop broken submissions from being accepted.

medler3• 3 years ago

Yes, I have seen that and this is completely utterly nonsense. I have absolutely no idea how you came to the conclusion that this would work.

I’m shocked someone would talk like this publicly from their company email. AMD doesn’t seem like a company I would want to work for.

blueflow• 3 years ago

If you are taking MR's from the public you are somewhat DDoS'ed with patches that are of low quality and will likely cause problems in your codebase. I totally understand that not everyone has the patience to deal with that with tactful language.

Not implying that this mail must be an example of this.

egberts12• 3 years ago

LOL!

Man, my former company could have chipped this device-memory mapped TCP stack for Linux kernel 2.14 to 2.24.

I created that. And it likely going to disappear forever although the product may still being in use today.

What a waste.

Was coded primarily to recapture the CPU burn on lower-end Pentium platform.

Good stuff, (sigh)

It had a project name but still unclear as to my contractual obligation with regard to reveal anything else about it (other than resume stuffing).

Linux Code for “Device Memory TCP” – Network to/from Accelerator RAM

Related Stories