Network Aware Application Mobility (NAAM) Charter
Goals
Data Centers offer application services to end-users such as video gaming, cloud
computing and others. As the application servers are distributed geographically in many
Data Centers for various reasons, the decision as where to instantiate services needs to be
optimized across end-user experience and data center economy.
User experience is impacted by Intra Data Center factors such as server load conditions,
and connectivity within Data Center as well as the underlying network behavior (such as
congestion, jitter, latency and loss) between end-user and the hosting Data Center and
between Data Centers.
Without optimization and network-awareness, services such as video gaming,
conferencing, on-demand video/HDTV applications, cloud computing work may receive
insufficient network resources (bandwidth, latency, reliability) or application sources.
The application stratum needs to be able to distribute applications to provide the services
to the user while balancing load on the network and processing power.
Applications hosted by Data Centers utilize proprietary means to make decision as where
to:
1) instantiate the application among the existing server locations (server selection
problem); or
2) place new applications or servers when existing servers cannot meet the service
request (application mobility problem).
The applications have two issues: a) no ability to query the network, and b) the lack of a
standard protocol to allow colluding of applications across domains.
The Working Group will design and specify:
1) the application-network communication protocol that allows the application to
query the whole network (Network Stratum query), and
2) application-to-application protocol that allows applications to cooperate in placing
new applications and servers (Server-placement colluding) to optimize application
mobility.
3) inter-provider consideration that allows negotiation between Data Centers
concerning charge for application, capability of Data Centers, etc and between
Data Centers and network carriers concerning network information sharing and
privilege, etc.
Why needed:
Due to the lack of cross stratum interaction between the application stratum and the
underlying network stratum during application service provisioning, application services
may make poor use of network resources or not achieve their overall quality of service
objectives and/or being denied of service provisioning pre-maturely.
This working group will focus on the Data Center applications (first). A second topic
will be Data Center Video to mobile nodes (after re-chartering).
Application Stratum Needs
The application CSO gateway (A-CSG) in the application stratum needs to collect
application level usage/performance data such as: (i) Max number simultaneous instances
of the application usage or some other measure of the application; (ii) Storage assignable
(e.g., max); (iii) Physical or Virtual assignment of processing, memory; (iv) Storage
access rate (disk, RAM, etc.), etc.
Whether virtual machine instances can be created in a different location and the
mechanics of moving virtual machines are of the need of the application stratum.
Whether an application must execute in multiple physical locations to provide diversity
and nearly instantaneous failover is another need of the application stratum.
The Working group will design and specify the application-to-application communication
interface/protocol and mechanisms that will enable to collect the application-level data
mentioned above.
In making decisions as where to instantiate the application among the existing server
locations or where to place new applications or servers when existing servers cannot meet
the service request, the ability to query the network is a critical to make the application
placement work effectively. The network query verifies if application instantiation works
or not.
The Working Group will design and specify the Application-Network communication
protocol and mechanisms that will enable coordinated query of application and network
requirements to available computing and network resources to determine the constraining
factor.
Objectives
The Working Group will focus on the following items:
-
A “problem statement” document that provides the description of problems and
key terminology.
-
An “architecture” document that provides the description of architecture
components and their functions. This document also provides how the network
query from application interacts with network processes to provide the data
requested data.
-
A “survey” document that lists all existing works in this area
-
A “requirement” document that specifies requirements for
the NAAM functionalities between Application Stratum and Network Stratum,
identifying types of messages and information CSO interface may need for
NAAM function.
-
A NAAM “protocol” document that defines request/reply format and semantics to
allow identified functionalities on the CSO interface based on the “requirement”
document.
-
A “protocol” document that defines interfaces between network layer controller
and internet NM and policy control (PEP/PC)
-
A “protocol” document that defines interfaces between network layer controller
and path computation element (PCE) for multi-domain path computation
-
An “implementation” experience document that describes CSO implementation
experience or network trial.
Proposed Schedule for IETF Working Group
Problem statement
Architecture statement
Survey Document
Requirements Document
AS-NS Protocol
NS-NM & PCP/PEP Protocol
NS-PCE
Implementation Report
[March 2011]
[March 2011]
[March 2011]
[July 2011]
[July 2011]
[July 2011]
[July 2011]
[November 2011]
Interested Parties:
Vendors:
- Huawei, Ericsson, Cisco, ETRI, Iridescent
Carriers:
- Telefonica, Verizon, NTT, KDDI, BT, DT, China Mobile
People (Attendees for Bar BOF) – 40 signed attendees
Young Lee
Dave McDysan
Katsushi Kobayashi
Xiangsong Cui
Dirk Hugo
Greg Bernstein
Matthias Kaschub
Aki Myrhinen
Heeyoung Jung
Dan Li
Xie Lei
Hideki Sunahara
Daniele Ceccarelli
Dave Croker
Myung Ki Shin
Dapeng Lin
Tae Yeon Kim
Oscar Gonzales de
Dios
Fatai Zhang
Mirja Kuhlewind
Quintin Zhao
Duan Shihui
Attila Takacs
Sandoche
Balakrichenan
Tomonori Takeda
Malcolm Betts
Jonathan Saddler
Yuji Tochio
Tian Linyi
Alvaro retana
Iguas Bagdonas
Guoying Zhang
Xu Yunbin
Antti Makela
Riccardo Bernadini
Kathy McEwen
Bob Briscoe
yajuan Wu
Yosifumi Atarashi
Ray Atarashi
ylee@huawei.com
dave.mcdysan@varizon.com
ikib@AI.AIST.GO.JP
Xiangsong.cui@huawei.com
dirk.vou-hugo@telecom.de
gregb@grotto-networking.com
kaschub@ikr.uni-stuttgart.de
anyrhine!@cs.helsinki.fi
hyjung@etri.re.kr
danli@huawei.com
xielei57471@huawei.com
suna@wide.ad.jp
daniele.ceccarelli@ericsson.com
dhc@dcrocker.net
mkshin@etri.re.kr
lindapeng@chinamobile.com
tykim@etri.re.kr
Huawei
Verizon
AIST
Huawei
DT
Grotto
Univ. Stuttgart
ogondio@tid.cs
zhangfatai@huawei.com
mirja.kuehlewind@ikr.uni-shuttgart.de
qzhao@huawei.com
duanshihui@mail.ritt.com.cn
attila.takacs@ericsson.com
Telefonica
Huawei
Univ. Stuttgart
Huawei
sandoche.balakrichenan@afnic.fr
takeda.tomonori@lab.ntt.co.jp
malcolm.betts@zte.com.cn
saddler@homesys.org
tochio@jp.fujitsu.com
tianlinyi@huawei.com
aretana@cisco.com
bagdona@cisco.com
zhangguoying@mail.ritt.com.cn
xuyunbin@catr.ca
antti.makela@tlek.fi
riccardo.bernardini@unind.it
k.mccwen@verizon.net
bob.briscoe@bt.com
wujajuan@huawei.com
ETRI
Huawei
Huawei
Ericsson
ETRI
China Mobil
ETRI
Ericsson
NTT
ZTE
Tellabs
Fujitsu
Huawei
Cisco
Cisco
Iridescent
BT
Huawei
Alaxala
IIJ
Schedule for creation of material for IETF BOF’s consideration:
Preliminary Problem statement
Preliminary Architecture statement
Preliminary Survey Document
Preliminary Requirements Document
Preliminary AS-NS Protocol
Preliminary NS-NM & PCP/PEP Protocol
Preliminary NS-PCE
Preliminary Implementation Report
[September 2010 ]
[September 2010 ]
[September 2010 ]
[October 2010]
[October 2010]
[October 2010]
[October 2010]
[November 2011]
Initial BOF work schedule
Interim Meetings – 3 Meetings (September– October)
Potential collaboration: NG Virtual networks, Internet 2, ITRI.