Internet DRAFT - draft-pskim-banana-mobilenetwork-mir
draft-pskim-banana-mobilenetwork-mir
Network Working Group P. Kim
Internet-Draft Korea Polytechnic University
Intended status: Experimental
Expires: January 1, 2019
July 01, 2018
A Bandwidth Aggregation Scheme on Mobile Network
with Multi-Interfaced Router
draft-pskim-banana-mobilenetwork-mir-00
Abstract
This draft considers a packet distribution scheme for bandwidth
aggregation on the mobile network with a multi-interfaced router
(MIR). In the proposed scheme, the MIR with multiple heterogeneous
wireless network interfaces effectively and fairly distributes
packets over end-to-end multi-path through multiple network
interfaces. Each network interface is considered to have a
distribution counter associated with corresponding end-to-end path.
This distribution counter varied by both weighted capacity and
distributed packets is used to determine if a network interface has
enough credits to distribute incoming packets on corresponding
end-to-end path. As a useful design parameter, the capacity unit can
be shown to make the performance of the proposed scheme as good as
possible.
Status of This Memo
This Internet-Draft is submitted in full conformance with the
provisions of BCP 78 and BCP 79.
Internet-Drafts are working documents of the Internet Engineering
Task Force (IETF). Note that other groups may also distribute
working documents as Internet-Drafts. The list of current Internet-
Drafts is at https://datatracker.ietf.org/drafts/current/.
Internet-Drafts are draft documents valid for a maximum of six months
and may be updated, replaced, or obsoleted by other documents at any
time. It is inappropriate to use Internet-Drafts as reference
material or to cite them other than as "work in progress."
This Internet-Draft will expire on January 1, 2019.
Kim Expires January 2, 2019 [Page 1]
Internet-Draft Bandwidth Aggregation Scheme July 2018
Copyright Notice
Copyright (c) 2018 IETF Trust and the persons identified as the
document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal
Provisions Relating to IETF Documents
(https://trustee.ietf.org/license-info) in effect on the date of
publication of this document. Please review these documents
carefully, as they describe your rights and restrictions with respect
to this document. Code Components extracted from this document must
include Simplified BSD License text as described in Section 4.e of
the Trust Legal Provisions and are provided without warranty as
described in the Simplified BSD License.
Table of Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2
2. Proposed Scheme . . . . . . . . . . . . . . . . . . . . . . . 3
2.1 Design Parameters . . . . . . . . . . . . . . . . . . . . . . 3
2.2 Operation Procedure and Example . . . . . . . . . . . . . . . 4
2.3 Performance Indices and Useful Design Parameters. . . . . . . 5
3. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 6
4. References . . . . . . . . . . . . . . . . . . . . . . . . . 6
4.1. Normative References . . . . . . . . . . . . . . . . . . 6
4.2. Informative References . . . . . . . . . . . . . . . . . 7
Author's Address . . . . . . . . . . . . . . . . . . . . . . . . 7
1. Introduction
The BANdwidth Aggregation for Network Access (BANANA) BoF was
chartered to develop solution(s) to support dynamic path selection
on a per-packet basis in networks that have more than one point of
attachment to the Internet. Bandwidth Aggregation consists of
splitting local traffic across multiple Internet links on a
per-packet basis, including the ability to split a single flow
across multiple links when necessary[1][2].
This draft considers the mobile network with a multi-interfaced
router (MIR)[3][4]. In addition, to consider the heterogeneous
wireless network environment, the MIR can be assumed to have
multiple heterogeneous wireless network interfaces. Therefore, the
MIR establishes simultaneously multiple paths to the Internet
through external wireless interfaces such as wireless local area
network (WLAN) and wireless wide area network (WWAN) with high
mobility and wide coverage. However, due to bandwidth constraints of
multi-path through external wireless interfaces, the MIR might
Kim Expires January 1, 2019 [Page 2]
Internet-Draft Bandwidth Aggregation Scheme July 2018
require a bandwidth aggregation to get sufficient bandwidth for MNs'
demanding inside a mobile network. As shown in [5], the bandwidth
aggregation requires generally several functions such as bandwidth
estimation, packet distribution, packet reordering, etc. Among them,
this draft focuses on the packet distribution scheme which
effectively and fairly distributes packets on the appropriate
end-to-end path through the corresponding network interface.
Therefore, this draft proposes a packet distribution scheme on the
MIR with heterogeneous wireless network interfaces for mobile
networks. Since the MIR is likely to have limited resources compared
with a general router, the proposed scheme adopts the frame-based
behavior that has lower complexity than the priority-based behavior.
In the proposed scheme, the MIR with multiple heterogeneous wireless
network interfaces effectively and fairly distributes packets over
end-to-end multi-path. Each network interface is considered to have
a distribution counter associated with the corresponding end-to-end
path. This distribution counter is used to determine if a network
interface has enough capacity to distribute packets on the
corresponding end-to-end path. The distribution counter can get
credits by the weighted capacity in bytes. The weighed capacity is
operated at the byte level and is added more to the distribution
counter with higher weight than that with less weight. On the other
hand, the distribution counter is decreased by the size of packets
being distributed. Thus, the distribution counter for each network
interface is varied by distributed packets as well as weighted
capacity.
In the proposed scheme, performance indices can be defined by ratio
and amount of distributed packets, packet loss, and throughput. The
capacity unit is shown to be a useful design parameter to make the
performance of the proposed scheme as good as possible.
2. Proposed Scheme
This draft considers the mobile network where the MIR has multiple
heterogeneous wireless network interfaces. The MIR establishes
multiple communication paths to the Internet through external
wireless interfaces such as WLAN and WWAN with high mobility and
wide coverage. In this mobile network environment, a fair packet
distribution scheme on the MIR with heterogeneous wireless network
interfaces is proposed for mobile networks.
2.1 Design Parameters
The MIR distributes packets effectively and fairly on the appropriate
end-to-end path through the corresponding network interface. Each
network interface is considered to have a distribution counter
associated with the corresponding end-to-end path. This distribution
Kim Expires January 1, 2019 [Page 3]
Internet-Draft Bandwidth Aggregation Scheme July 2018
counter is used to determine if a network interface has enough
capacity to distribute packets on the corresponding end-to-end path.
The distribution counter can get credits by the weighted capacity in
bytes. The weighted capacity is defined by Weighted capacity =
Capacity unit*Weight. The capacity unit in bytes is a useful design
parameter and thus can affect on the performance of the proposed
scheme. The weight is determined proportionately from the estimated
available bandwidth of end-to-end paths. The weighed capacity is
operated at the byte level and is added more to the distribution
counter with higher weight than that with less weight. On the other
hand, the distribution counter is also decreased by the size of
packets being distributed. Thus, the distribution counter for each
network interface is varied by distributed packets as well as
weighted capacity.
2.2 Operation Procedure and Example
The operation procedure for each round is as follows. For the first
path, packets are distributed when the distribution counter is
greater than the incoming packet's size. If it is lower, the
distribution counter is increased by the weighted capacity and then
the incoming packet is distributed on the current path. Then the
distribution counter is decreased by the size of packet being
distributed. If the distribution counter is still lower than the
incoming packet's size, the incoming packet held back until the
proposed scheme moves on the next path. After visiting all paths,
the round is finished. The above operation procedure in next round
is repeated when there are incoming packets.
As an example, the MIR is assumed to have three heterogeneous
wireless network interfaces and thus there are three communication
paths. These paths are called the Green (high bandwidth), Yellow
(medium bandwidth), Red (low bandwidth) paths, respectively. Since
this paper focuses on the packet distribution scheme, available
bandwidths for three paths through corresponding interfaces are
assumed to have fixed weight ratio 4:2:1. The capacity unit is set
by 256 bytes and thus these paths have weighted capacity 1024, 512,
256 bytes, respectively. All initial values of distribution counters,
denoted by DCg,y,r, and total amount of distributed packets, denoted
by ADPg,y,r, for three paths are set with 0. There are four kinds
of packet type with different sizes such as 256, 512, 768, 1024
bytes.
Incoming packets are waiting to be distributed to the most
appropriate path.
A(1024) <- B(256) <- C(768) <- D(512) <- E(768)
Kim Expires January 1, 2019 [Page 4]
Internet-Draft Bandwidth Aggregation Scheme July 2018
The 1st round is operated. At Green path, the weighted capacity is
added and thus the distribution counter, DCg, is 1024. Then, since
the DCg is not less than the incoming packet's size (1024 bytes), the
incoming packet 'A' is distributed to Green path and thus DCg=0.
Currently, total amount of distributed packets to Green path, ADPg,
is 1024 bytes. At Yellow path, the weighted capacity is added and
thus DCy=512. Then, since the DCy is not less than the incoming
packet's size (256 bytes), the incoming packet 'B' is distributed to
Yellow path and thus DCy=256. Currently, ADPy=256. At Red path, the
weighted capacity is added and thus DCr=256. Then, since the DCr is
still less than the incoming packet's size (768 bytes), move to Green
path. Currently, ADPr=0. The 1st round is done. Then, the 2nd round
is operated. At Green path, since the DCg is less than the incoming
packet's size (768 bytes), the weighted capacity is added and thus
DCg=1024. Then, the incoming packet 'C' is distributed to Green path
and thus DCg=256. Currently, ADPg=1792. At Yellow path, since DCy is
less than the incoming packet's size (512 bytes), the weighted
capacity is added and thus DCy=768. Then, the incoming packet 'D' is
distributed to Yellow path and thus DCy=256. Currently, ADP =768. At
Red path, since DCr is less than the incoming packet's size (768
bytes), the weighted capacity is added and thus DCr=512. Then, the
incoming packet 'E' is distributed to Green path and thus DCg=0.
Currently, ADPr=512. The 2nd round is done. The operation procedure
in next round is repeated when there are incoming packets.
2.3 Performance Indices and Useful Design Parameters
There can be four performance indices in the proposed scheme; ratio
of distributed packets, amount of distributed packets, packet loss
and throughput. Of course, the throughput can be improved as many
packets are distributed. However, the improvement of the throughput
does not have the meaning if the packet loss increases.
Performance Index Objective
--------------------------------------------------------------------
Ratio of distributed packets Distributing packets fairly according
to the weight ratio of each
end-to-end path
Amount of distributed packets Distributing packets as many as
possible over each end-to-end path
Packet loss Minimizing lost packets on each
end-to-end path
Throughput Maximizing the average rate of
successful packet delivery on each
end-to-end path
--------------------------------------------------------------------
Kim Expires January 1, 2019 [Page 5]
Internet-Draft Bandwidth Aggregation Scheme July 2018
As mentioned before, the capacity unit is a useful design parameter
to determine weighted capacity that affect on four performance
indices mentioned before. Too big value of the capacity unit can
introduce excessive credits for end-to-end paths, which means that
network paths have enough credits to distribute packets. Thus,
incoming packets are more likely to be distributed simultaneously on
every path each round, which cannot provide fair distribution
according to weights for network paths. Thus, the ratio of
distributed packets can be degraded. In addition, since packets can
be distributed too many over each end-to-end path, the performance
for packet loss can be degraded. Of course, since packets are
distributed too many over each end-to-end path, the throughput can
be improved. However, as mentioned before, the improvement of the
throughput does not have the meaning since the packet loss
increases. On the other hand, too small value of the capacity unit
can introduce deficient credits for end-to-end paths, which means
that end-to-end paths do not have enough credits to distribute
packets. Thus, incoming packets are less likely to be distributed on
every paths each round, which can thus degrade the amount of
distributed packets. In addition, since the amount of distributed
packets over each end-to-end path is not much, the throughput can be
degraded whereas the packet loss can decrease. Therefore, the
important issue here is how to choose an appropriate capacity unit
to make the performance of the proposed scheme as good as possible.
3. IANA Considerations
This document has no IANA actions.
4. References
4.1. Normative References
[1] N. Leymann, C. Heidemann, et al, "BANdwidth Aggregation for
interNet Access (BANANA) The Control Protocol of Bonding
Tunnels", draft-leymann-banana-signaling-02, work in progress.
[2] N. Leymann, C. Heidemann, et al, "BANdwidth Aggregation for
interNet Access (BANANA) The Data Plane of Bonding
Tunnels", draft-leymann-banana-data-encap, work in progress.
[3] M. Blanchet and P. Seite, "Multiple Interfaces and Provisioning
Domains Problem Statement", IETF RFC 6418, November 2011.
[4] M. Wasserman and P. Seite, "Current Practices for
Multiple-Interface Hosts", IETF RFC 6419, November 2011.
Kim Expires January 2, 2019 [Page 6]
Internet-Draft Bandwidth Aggregation Scheme July 2018
4.2. Informative References
[5] K. Habaka, K. A. Harras, M. Youssef, "Bandwidth aggregation
techniques in heterogeneous multi-homed devices : A survey",
Computer Networks, vol. 92, pp. 168~188, 2015.
Author's Address
Pyungsoo Kim
Department of Electronics Engineering,
Korea Polytechnic University,
2121 Jungwang-Dong, Shiheung City,
Gyeonggi-Do 429-793
KOREA
Phone: +82 31 8041 0489
EMail: pskim@kpu.ac.kr
Kim Expires January 1, 2019 [Page 7]
