NUSE talk at #osio http://operatingsystems.io/

1. Network Stack in
Userspace (NUSE)
!
Hajime Tazaki
Ryo Nakamura
(University of Tokyo)
!
New Directions in Operating Systems
London, 2014

2. Motivation
Implementation of the Internet
is not finished yet
!
!
Faster evolution of OSes (network
stack)
OS personalization
2

3. I have a new Layer-3/4
protocol! Yay!
I have new, great Layer-3/4 protocol ! It
will change the WORLD !
Replace network stack ?
No: destroy my life ?!
(experimental ? not tested ?)
Yes: I wanna be your slave.
Slow evolution of network stack ?
VM on personal device ?
3

4. Virtual Machine ?
Poll: “When you download and run software, how often do you use a virtual machine (to reduce
security risks)?”
Jon Howell, Galen Hunt, David Molnar, and Donald E. Porter, Living Dangerously: A Survey of Software Download
Practices, no. MSR-TR-2010-51, May 2010
4

5. costin.raiciu@cs.pub.ro, j.araujo@ucl.ac.uk, rizzo@iet.unipi.it
Internet paths
that it is still
despite the
the blame
extensions taking
placed on end
moving protocols
deployment
optimizations.
support for user-level
commodity
number of
host stack,
s.
our mux/de-mux
line rate (up
Slow evolution of network stack
Honda et al., Rekindling Network Protocol Innovation with User-Level Stacks, ACM
SIGCOMM CCR, Vol.44, Num. 2, April 2014
cores, and
over a basic
same server
1.00
0.75
0.50
0.25
0.00
2007 2008 2009 2010 2011 2012
Date
Ratio of flows
Option
SACK
Timestamp
Windowscale
Direction
Inbound
Outbound
Figure 1: TCP options deployment over time.
pen infrequently not only because of slow release cycles, but
also due to their cost and potential disruption to existing
setups. If protocol stacks were embedded into applications,
they could be updated on a case-by-case basis, and deploy-ment
would be a lot more timely.
For example, Mac OS, Windows XP and FreeBSD still
use a traditional Additive Increase Multiplicative Decrease
(AIMD) algorithm for TCP congestion control, while Linux

6. Meanwhile in
Filesystem world..
There is,
Filesystem in Userspace
(FUSE)
Userspace code can host
new filesystem (sshfs,
GmailFS, etc)
Performance is bad,
but doesn’t matter
Flexibility and
functionality do matter
6
http://fuse.sourceforge.net/

7. Alternatives
Container (LXC, OpenVZ, vimage)
share kernel with host operating system (no
flexibility)
Library OS
full scratch: mtcp, Mirage, lwIP
Porting: OSv, Sandstorm, libuinet (FreeBSD),
Arrakis (lwIP), OpenOnload (lwIP?)
Glue-layer: LKL (Linux-2.6), rumpkernel (NetBSD)
7

8. Network Stack
in Userspace

9. What’s NUSE ?
Network stack in Userspace
A library operating system
Library version of network
stack (of monolithic kernel)
Linux (latest), FreeBSD (plan)
(UNIX) Process-based
virtualization
9
nuse example
kernel bypassed
TCP/IP
ARP/
ndisc
libnuse
glibc
NIC
userspace
kernel
raw sock
netmap
DPDK (etc)

10. Why NUSE ?
minimized porting effort
Linux (net-next) changes frequently
!
full functional network stack for
netmap
DPDK
(any kernel-bypass technology)
10

11. How it works
Application
POSIX glue
TCP UDP DCCP SCTP
ICMP ARP
IPv6 IPv4
Qdisc
Netfilter Bridging
Netlink
IPSec Tunneling
Kernel layer
NUSE core
bottom halves/
rcu/timer/
interrupt
struct
net_device
RAW DPDK netmap ...
NIC
petit-scheduler
1. (monolithic) kernel
source
2. scheduler
3. POSIX glue
redirect system calls
4. network I/O
raw socket, DPDK,
netmap, etc..
11

12. 1) kernel build
Application
POSIX glue
TCP UDP DCCP SCTP
ICMP ARP
IPv6 IPv4
Qdisc
Netfilter Bridging
Netlink
IPSec Tunneling
Kernel layer
NUSE core
bottom halves/
rcu/timer/
interrupt
struct
net_device
RAW DPDK netmap ...
NIC
petit-scheduler
patch to kernel tree
with new (hw independent)
arch (arch/sim)
robust to (frequent)
mainstream changes
12

13. 2) scheduler
Application
POSIX glue
TCP UDP DCCP SCTP
ICMP ARP
IPv6 IPv4
Qdisc
Netfilter Bridging
Netlink
IPSec Tunneling
Kernel layer
NUSE core
bottom halves/
rcu/timer/
interrupt
struct
net_device
RAW DPDK netmap ...
NIC
petit-scheduler
offer alternate
context primitives
interrupts, timer,
thread, bottom halves
(tasklet, workqueue,
waiter, etc)
wrap with POSIX thread
easily debuggable
ucontext fiber for low
overhead (not yet)
13

14. 3) POSIX glue code
Application
POSIX glue
TCP UDP DCCP SCTP
ICMP ARP
IPv6 IPv4
Qdisc
Netfilter Bridging
Netlink
IPSec Tunneling
Kernel layer
NUSE core
bottom halves/
rcu/timer/
interrupt
struct
net_device
RAW DPDK netmap ...
NIC
petit-scheduler
Hijack function calls
socket => nuse_socket
read => nuse_read
apps not aware of
LD_PRELOAD=libnuse.so ..
14

15. 4) network I/O
Application
POSIX glue
TCP UDP DCCP SCTP
ICMP ARP
IPv6 IPv4
Qdisc
Netfilter Bridging
Netlink
IPSec Tunneling
Kernel layer
NUSE core
bottom halves/
rcu/timer/
interrupt
struct
net_device
RAW DPDK netmap ...
NIC
petit-scheduler
connect NUSE to NIC
options
raw socket (default)
DPDK (if available)
netmap (if available)
Tap
15

16. Usage
download
git clone git://github.com/libos-nuse/net-next-
nuse
compile
make library ARCH=sim (NETMAP=yes) (DPDK=yes)
execute
sudo NUSECONF=nuse.conf ./nuse (application)
16

17. configs
17
# Interface definition.!
interface eth0!
address 192.168.0.10!
netmask 255.255.255.0!
macaddr 00:01:01:01:01:01!
viftype TAP!
!
interface p1p1!
address 172.16.0.1!
netmask 255.255.255.0!
macaddr 00:01:01:01:01:02!
!
# route entry definition.!
route!
network 0.0.0.0!
netmask 0.0.0.0!
gateway 192.168.0.1

18. (possible) use cases
New protocol deployment
Chrome + Linux mptcp (on NUSE)
Process-level virtual instance
% NUSE-linux-ovs | NUSE-freebsd-NAT |
NUSE-router | NUSE-nginx!
VM chaining via UNIX command line
18

19. Limitation (ongoings)
no fork(2)/exec(2) support
no multi-processes
no sysctl/proc
(inefficient) thread scheduling
19

20. Experiments
1. Can we benefit with OS personalization?
present a custom (NUSE) kernel with an
application (OS personalization)
2. How much overhead does NUSE add?
Simple performance measurements
20

21. Tested applications
working
ping, iperf, nginx (partially), sleep,
need patches
nc, wget, dig, host
21

22. Setup: Performance
measurement
NUSE node Tx/Rx nodes
CPU Xeon E5-2650v2 @
ping!
flowgen
10G 10G
22
2.60GHz
(16 core)
Xeon L3426 @ 1.87GHz
(8 core)
Memory 32GB 4GB
NIC Intel X520 Intel X520
ping!
flowgen
vnstat!
(packet count)
Tx NUSE Rx

23. Host Tx
(NUSE->Receiver)
NUSE Rx
23
avg max min
dpdk! 2.610 8.000 0.156
netmap 0.370 0.494 0.252
raw 0.396 0.501 0.290
tap 0.397 0.538 0.303
500
450
400
350
300
250
200
150
100
50
0
dpdk netmap raw tap
Throughput (Mbps)
ping (RTT) throughput
(1024byte,UDP)
8
7
6
5
4
3
2
1
0
dpdk netmap raw tap
RTT (ms)

24. L3 Routing
Sender->NUSE->Receiver
Tx NUSE Rx
24
avg max min
dpdk! 11.998 27.700 0.252
netmap 0.664 0.741 0.556
raw 0.663 0.761 0.575
tap 0.694 0.749 0.602
ping (RTT)
500
450
400
350
300
250
200
150
100
50
0
netmap raw tap
Throughput (Mbps)
throughput
(1024byte,UDP)
30
25
20
15
10
5
0
dpdk netmap raw tap
RTT (ms)

25. Discussions
not so bad performance
we don’t care much about performance
network stack is full functional
but supplemental tools are not sufficient
25

26. Network Simulator
Integration (ns-3)
network stack +ns-3 network simulator
!
Direct Code Execution (DCE)
Established by Mathieu Lacage (2006)
part of ns-3 project
!
Features
reproducible (deterministic clock)
controllable (simulator’s facility)
http://www.nsnam.org/overview/projects/direct-code-execution/
26

27. 27

28. NUSE vs DCE
NUSE DCE
shared
kernel library ARCH=sim ARCH=sim
scheduler (host) pthread simulator’s scheduler!
28
(deterministic)
POSIX hijack hijack
network I/O raw/netmap/DPDK/tap ns3:NetDevice
execution LD_PRELOAD
dlmopen(3)!
single proc/multi-instances

29. DCE Architecture
Application
(ip, iptables, quagga)
POSIX layer
TCP UDP DCCP SCTP
ICMP ARP
IPv6 IPv4
Netfilter Bridging
struct net_device
29
Qdisc
Netlink
IPSec Tunneling
bottom halves/rcu/
timer/interrupt
Kernel layer
Heap Stack
memory
Virtualization Core
layer
ns-3 (network simulation core)
DCE
ns-3
applicati
on
ns-3
TCP/IP
stack
3) POSIX!
Layer
1) Core!
Layer
2) Kernel!
Layer

30. Bug reproducibility
Home Agent
AP1 AP2
30
Wi-Fi Wi-Fi
handoff
ping6
correspondent
node
mobile node
(gdb) b mip6_mh_filter if dce_debug_nodeid()==0
Breakpoint 1 at 0x7ffff287c569: file net/ipv6/mip6.c, line 88.
<continue>
(gdb) bt 4
#0 mip6_mh_filter
(sk=0x7ffff7f69e10, skb=0x7ffff7cde8b0)
at net/ipv6/mip6.c:109
#1 0x00007ffff2831418 in ipv6_raw_deliver
(skb=0x7ffff7cde8b0, nexthdr=135)
at net/ipv6/raw.c:199
#2 0x00007ffff2831697 in raw6_local_deliver
(skb=0x7ffff7cde8b0, nexthdr=135)
at net/ipv6/raw.c:232
#3 0x00007ffff27e6068 in ip6_input_finish
(skb=0x7ffff7cde8b0)
at net/ipv6/ip6_input.c:197

31. Debugging
==5864== Memcheck, a memory error detector
==5864== Copyright (C) 2002-2009, and GNU GPL'd, by Julian Seward et al.
==5864== Using Valgrind-3.6.0.SVN and LibVEX; rerun with -h for copyright info
==5864== Command: ../build/bin/ns3test-dce-vdl --verbose
==5864==
==5864== Conditional jump or move depends on uninitialised value(s)
==5864== at 0x7D5AE32: tcp_parse_options (tcp_input.c:3782)
==5864== by 0x7D65DCB: tcp_check_req (tcp_minisocks.c:532)
==5864== by 0x7D63B09: tcp_v4_hnd_req (tcp_ipv4.c:1496)
==5864== by 0x7D63CB4: tcp_v4_do_rcv (tcp_ipv4.c:1576)
==5864== by 0x7D6439C: tcp_v4_rcv (tcp_ipv4.c:1696)
==5864== by 0x7D447CC: ip_local_deliver_finish (ip_input.c:226)
==5864== by 0x7D442E4: ip_rcv_finish (dst.h:318)
==5864== by 0x7D2313F: process_backlog (dev.c:3368)
==5864== by 0x7D23455: net_rx_action (dev.c:3526)
==5864== by 0x7CF2477: do_softirq (softirq.c:65)
==5864== by 0x7CF2544: softirq_task_function (softirq.c:21)
==5864== by 0x4FA2BE1: ns3::TaskManager::Trampoline(void*) (task-manager.==5864== Uninitialised value was created by a stack allocation
==5864== at 0x7D65B30: tcp_check_req (tcp_minisocks.c:522)
==5864==
Memory error detection
among distributed nodes
in a single process
using Valgrind
!
!
31

32. Fine-grained parameter coverage
Code coverage measurement with DCE
With fine-grained network, node, protocol parameters
32

33. Continuous integration
http://ns-3-dce.cloud.wide.ad.jp/jenkins/job/daily-net-next-sim/
33

34. Summary
NUSE (Network Stack in Userspace)
OS personalization (fast evolution, easy
deployment)
DCE (Direct Code Execution)
Flexible network experiment/test with
deterministic clock
34

35. github: https://github.com/libos-nuse/net-next-
nuse
DCE: http://bit.ly/ns-3-dce
twitter: @thehajime
35

36. Backups

37. Potentials of Userspace
Networking
High-performance networking
Useful debugging facilities
Operating system personalization
37

38. 1) kernel build
build kernel source tree w/ the patch
make menuconfig ARCH=sim
make library ARCH=sim
➔ libnuse-linux-3.17-rc1.so
38

39. Example: How timer
works
39
add_timer()
TIMER_SOFTIRQ
timer_list
run_timer_softirq ()
timer handler
timer thread
(timer_create (2))

40. 3) POSIX glue code
extern int sim_sock_socket (int,int,int, struct socket **);
int socket (int family, int type, int proto)
{
sim_update_jiffies ();
struct socket *kernel_socket =
sim_malloc (sizeof (struct socket));
memset (kernel_socket, 0, sizeof (struct socket));
int ret = sim_sock_socket (family, type, proto, &kernel_socket);
g_fd_table[curfd++] = kernel_socket;
sim_softirq_wakeup ();
return curfd - 1;
}
https://github.com/libos-nuse/net-next-nuse/blob/nuse/arch/sim/nuse-glue.c
40

41. Tx callgraph
sendmsg () (socket API)
sim_sock_sendmsg () (NUSE)
sock_sendmsg ()
ip_send_skb ()
ip_finish_output2 ()
dst_neigh_output () (ex-kernel)
neigh_resolve_output ()
arp_solicit ()
dev_queue_xmit ()
sim_dev_xmit () (NUSE)
nuse_vif_raw_write ()
41

42. Rx callgraph
start_thread () (pthread)
nuse_netdev_rx_trampoline ()
nuse_vif_raw_read () (NUSE)
sim_dev_rx ()
netif_rx () (ex-kernel)
start_thread () (pthread)
do_softirq () (NUSE)
net_rx_action ()
process_backlog () (ex-kernel)
__netif_receive_skb_core ()
ip_rcv ()
42
vNIC!
rx
softirq!
rx