One document matched: draft-cheng-supa-ddc-use-cases-06.txt
Differences from draft-cheng-supa-ddc-use-cases-05.txt
Network Working Group Y. Cheng
Internet-Draft China Unicom
Intended status: Informational JF. Tremblay
Expires: October 5, 2015 Viagenie
J. Bi
Tsinghua University
L. M. Contreras
Telefonica I+D
April 3, 2015
Use Case for Distributed Data Center in SUPA
draft-cheng-supa-ddc-use-cases-06
Abstract
Large scale DCs can provide various services and usually have a lot
of internal and external links, and these links are usually VPNs.
The Service provisioning and network connectivity configurations are
complex and sometimes dynamic, for which manual configuration is not
suitable. This draft analyzes the use case in Distributed Data
Centers (DDC), in which some VPN scenarios are covered, and the
applicability of Simplified Use of Policy Abstractions (SUPA) data
models which can be used for better and automated resource usage and
easy service/network configuration.
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 http://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 October 4, 2015.
Copyright Notice
Copyright (c) 2015 IETF Trust and the persons identified as the
document authors. All rights reserved.
Cheng, et al. Expires October 4, 2015 [Page 1]
Internet-Draft Use cases for DDC Applications in SUPA April 2015
This document is subject to BCP 78 and the IETF Trust's Legal
Provisions Relating to IETF Documents
(http://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. Requirements Language . . . . . . . . . . . . . . . . . . . . 3
3. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 3
4. Challenges Faced by Data Center ISPs . . . . . . . . . . . . 4
5. SUPA Benefits . . . . . . . . . . . . . . . . . . . . . . . . 4
6. Scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . 5
6.1. Scenario:Inter DC Connectivity . . . . . . . . . . . . . 5
6.2. Scenario:vDC Connectivity . . . . . . . . . . . . . . . . 7
6.3. Scenario:Dynamic Link Configuration for DC . . . . . . . 9
6.4. Scenario:DC Connectivity for Virtual Private Clouds (VPC) 10
7. Security Considerations . . . . . . . . . . . . . . . . . . . 12
8. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 12
9. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 12
10. References . . . . . . . . . . . . . . . . . . . . . . . . . 12
10.1. Normative References . . . . . . . . . . . . . . . . . . 12
10.2. informative References . . . . . . . . . . . . . . . . . 12
Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 13
1. Introduction
The SUPA (Simplified Use of Policy Abstractions) work aims at
providing data models, including topology data models, network
service specific data models, policy data models, to easily,
accurately, and efficiently select and use the available
communication network capabilities. An example of the data model can
be found in [I-D.zaalouk-supa-configuration-model]. Service Manager
(SM) is used by an a communications service provider and/or operator
to deploy and manage services on top of network facilities. An
example of SM is a set of applications used by an Operational Support
System (OSS) entity to perform network configuration. Several SUPA
use cases have been introduced in the problem statement document.
This document reviews various scenarios for Distributed Data Center
(DDC) use case.
Take a large-scale Distributed Data Center (DDC) operator as an
example, it provides server hosting, leased line, value-added
Cheng, et al. Expires October 4, 2015 [Page 2]
Internet-Draft Use cases for DDC Applications in SUPA April 2015
services to enterprises and ISPs, and has more than 10 data centers
using over one Tbps of bandwidth in a capital city. In this IDC
network, traffic at each site is routed via configuring policy routes
and adjusting routes prioritization to choose an outgoing link. This
type of static provisioning comes with high costs and poor
operability. Furthermore, the link bandwidth resources in the data
centers are not efficiently utilized.
In quite some of the scenarios, the links between DCs are VPNs,
including L2VPN, L3VPN, etc. SUPA will be mainly used for those VPN
configurations. Although there may be some cases where physical
links are used, but those are out of the scope of this draft.
DC and network may belong to different operators. If a DC operator
needs to configure network connectivity for DCs, it may need to
cooperate with network operators providing such connectivity.
Network operators can define and provide data models to enable this.
This document illustrates several distributed datacenter (DDC)
applications and explains how an operator could use SUPA to provide
these applications.
2. Requirements Language
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
document are to be interpreted as described in [RFC2119].
3. Terminology
The terminology used in the SUPA problem statement draft
[I-D.zhou-supa-framework] and
[I-D.karagiannis-supa-problem-statement] apply also to this draft.
DC Data Center
DDC Distributed Data Center
NM/NC Network Manager / Controller
OSS Operational Support System
SM Service Manager
SUPA Simplified Use of Policy Abstractions
TTM Time to Market
Cheng, et al. Expires October 4, 2015 [Page 3]
Internet-Draft Use cases for DDC Applications in SUPA April 2015
VAS Value Added Service
vDC virtual Data Center
VPC Virtual Private Cloud (PC)
4. Challenges Faced by Data Center ISPs
There are many challenges in traditional data centers:
1) infrastructure and network link is usually leased, depending on
manual planning and design, which leads to low resource usage and
high cost. In consequence, the operator that rent these resources
has to offer SUPA data models for facilitating control of them (for
instance, by the DC operator).
2) Service expansion is limited in a single physical DC. Each DC
resource is isolated, so service and resource can only be deployed in
one single DC.
3) VAS (Value Added Service) is provisioned via static configuration,
which brings complex training, long service TTM time and poor
flexibility. This could not meet the requirements of complex use
cases, e.g., lot of VAS devices, significant differences between
various services.
5. SUPA Benefits
In quite some cases, DC oprators need to optimize and automate
service deployment for them serlfves or for customers. While in some
cases, DC tenants also need to perform some optimization, e.g. a vDC
tennant may want traffic steering to make full use of links. To
solve the above challenges for data center oerators and tenants, SUPA
could be applied in the following ways:
o SUPA supports an open network architecture: standardizaed data
models enable an open architecture and make it possible for
unified service / network planning, which can interconnect with
third party cloud platform, supporting fast service innovation.
o SUPA supports overall DC resource integration: SUPA data models
can be used for network resource virtualization; inter-DC
resource, virtual DC (vDC) resource, etc, can be integrated and
controlled by a centralized functional entity.
o SUPA supports automatic E2E service delivery: Network (including
virtual network), computing, inter-DC management of storage
Cheng, et al. Expires October 4, 2015 [Page 4]
Internet-Draft Use cases for DDC Applications in SUPA April 2015
resource, automatic service delivery, automatic VPN connection
configurations between DCs, which improves operation efficiency.
o SUPA contributes to improve DDC network usage by means of
Intelligent scheduling of DDC traffic, improving link usage.
o SUPA supports VAS service on-demand provisioning automatically:
Create or delete VAS nodes on-demand, based on various service
requirements; network forwarding policy based on the VAS routing,
to achieve automatic draining and automatic configuration of VAS
device policy.
Please refer to [I-D.zhou-supa-framework] and other SUPA related
documents for more details of SUPA features.
The following sections will illustrate three typical cases in
distributed data center which could benefit from SUPA architecture.
6. Scenarios
In the following uses, Service Manager (SM) is used for service and
policy definition; and Network Manager (Controller) is used for
network topology maintenance and mapping data models to detail
network configruations, as defined in [I-D.zhou-supa-framework].
6.1. Scenario:Inter DC Connectivity
Cheng, et al. Expires October 4, 2015 [Page 5]
Internet-Draft Use cases for DDC Applications in SUPA April 2015
+---------------------------+
| Service Manager |
| |
| +----------+ +----------+ |
| |Service | |Policy | |
| |Data Model| |Data Model| |
| +----------+ +----------+ |
| |
+---------------------------+
^
|
|
|
v
+--------------------------+
| Network |
| Manager / Controller |
+--------------------------+
/ \
/ \
+---------+ \
/| DC1 |\ \
/ +---------+ \ +-----------+
| | d1 \___a1________| DC-A |
| | | |
| +---------+ +-----------+
d3 | | DC2 |\
| +---------+ \ +-----------+
| | d2 \___a2________| DC-B |
| | ____a3________| |
\ +---------+ / +-----------+
\| DC3 |/
+---------+
Scenario:Inter DC Connectivity
There can be a lot of links between data centers. Configuration of
these links is complex. As shown in Figure 1, service data models
and policy data models can be defined to automate the configuration
procedures. The service data model for connectivity will specify
attributes of (virtual) links, e.g. the end points of links,
bandwidth, QoS and availability parameters, etc. The policy model
can specify some high level requirements to the links, like routing
strategy (via and not via) and price/cost strategy. The policy data
model can also define the policy rules that drive the security
requirements.
Cheng, et al. Expires October 4, 2015 [Page 6]
Internet-Draft Use cases for DDC Applications in SUPA April 2015
Inter DC connections can be classified into two types: connections
within a single administrative domain and connections across multiple
administrative domains. Links d1, d2 and d3 are within an
administrative domain; and links a1, a2 and 3 are across domains.
The difference between them is that connections across multiple
administrative domain require extra negotiation and authentication/
authorization, which can be achieved via communications between SMs.
Data models for this purpose should also be defined.
The links interconnecting two DCs together should guarantee a minimum
bandwidth, certain QoS parameters, and provide availability
guarantees. As a service policy example in Figure 1, for traffic
from DC2 to DC-B, if the load on a link exceeds a threshold (e.g.,
90%), some (new) traffic can be redirect to another link.
6.2. Scenario:vDC Connectivity
+---------------------------+
| Service Manager |
| |
| +----------+ +----------+ |
| |Service | |Policy | |
| |Data Model| |Data Model| |
| +----------+ +----------+ |
| |
+---------------------------+
^
|
|
|
v
+--------------------------+
| Network |
| Manager / Controller |
+--------------------------+
/ | \
/ | \
/ | \
/ +-------------------+ \
/ | DC2 | \
/ | +---------------+ | \
/ | |Tenant1 (vDC) | | \
/ | +---------------+ | \
/ | | \
/ | +---------------+ | \
/ | | Tenantn (vDC) | | \
| | +---------------+ | |
| +-------------------+ |
Cheng, et al. Expires October 4, 2015 [Page 7]
Internet-Draft Use cases for DDC Applications in SUPA April 2015
| |vDC link |
| +-------------+ |
| | | |
| /| Cloud |\ |
| / +-------------+ \ |
| vDC link / \vDC link|
| / \ |
+-------------------+ +-------------------+
| DC1 | | DC3 |
| +---------------+ | | +---------------+ |
| | Tenant1 (vDC) | | | | Tenant1 (vDC) | |
| +---------------+ | | +---------------+ |
| | | |
| +---------------+ | | +---------------+ |
| | Tenantk (VDC) | | | | Tenantn (vDC) | |
| +---------------+ | | +---------------+ |
+-------------------+ +-------------------+
Scenario:vDC Connectivity
A DC tenant may have resources, e.g. network, computing, storage,
etc, in multiple physical DCs. DC operators will provide internal
network connectivity for these distributed resources, and make it
look like one seamless entity, which can be called as virtual DC
(vDC).
The internal links for vDC can be implemented via tunneling overlay
technologies, e.g. VPN or VxLAN, etc. The tunnels need to be
dynamically established, managed and released.
As show in Figure 2, service data model and policy data model can be
defined to automate the links configuration for vDCs. A policy model
should specify the attributes of the tunnels, e.g., bandwidth, QoS
and availability parameters, the interaction with policy systems that
dynamically scale the DC resources assigned to a tenant, and the
policy rules that drive the prioritization of resource assignments.
The networking resources assigned to a tenant should scale
proportionally to the compute resources assigned to a tenant. The
traffic should be prioritized to resources owned by tenants that
offer interactive services according to the time zone the DC is
located in.
Cheng, et al. Expires October 4, 2015 [Page 8]
Internet-Draft Use cases for DDC Applications in SUPA April 2015
6.3. Scenario:Dynamic Link Configuration for DC
Static and over provisioning for DC links is not always a good
solution. Sometimes dynamic configuration is necessary.
+---------------------------+
| Service Manager |
| |
| +----------+ +----------+ |
| |Service | |Policy | |
| |Data Model| |Data Model| |
| +----------+ +----------+ |
| |
+---------------------------+
^
|
|
|
v
+--------------------------+
| Network |
| Manager / Controller |______________
+--------------------------+ \
/ \ \
/ \ \
/ \ |
+--------------+ +-------------------+ |
| | | | |
| | | DC2 |
| |--------------- | | |
| DC1 | +-------------------+ |
| | |
| | \ |
| | \ +-------------------+ |
| | \ | | /
| | \__________| DC2 |/
+--------------| | |
+-------------------+
Scenario:Dynamic Link Configuration for DC
One case is virtual machine migration and large amount of data
transfer between DCs. But this kind of activity does not happen
frequently. A dedicated link with constant bandwidth for this
purpose is too expensive. The network operator should allow the DC
operator to create a link on demand when necessary. This link may
Cheng, et al. Expires October 4, 2015 [Page 9]
Internet-Draft Use cases for DDC Applications in SUPA April 2015
have large bandwidth but last for a limited time period. An
alternative is to create short-term dedicated links for backups and
migrations.
As shown in figure 3, data models can help to automate these kind of
configurations. In the data models, the attributes of links
(bandwidth, QoS and availability parameters) should be specified.
The policy concerning strict and soft bounds on the lifetime of such
links, and the policy concerning the scheduling of dedicated links
(e.g., based on the current load) and the services using the
dedicated links can also be specified.
When the traffic volume between DCs exceeds a certain threshold, the
policy-driven service manager requests that traffic schedules may be
adjusted within bounds in order to balance load on the links (e.g.,
delay backups and migrations until the network has the necessary
capacity).
6.4. Scenario:DC Connectivity for Virtual Private Clouds (VPC)
Cheng, et al. Expires October 4, 2015 [Page 10]
Internet-Draft Use cases for DDC Applications in SUPA April 2015
+---------------------------+
| Service Manager |
| |
| +----------+ +----------+ |
| |Service | |Policy | |
| |Data Model| |Data Model| |
| +----------+ +----------+ |
| |
+---------------------------+
^
|
|
|
v
+--------------------------+
| Network |
| Manager / Controller |______________
+--------------------------+ \
/ \ \
/ \ \
/ \ |
+-------------------+ +-------------------+ |
| Cloud for VPCs | | | |
| +---------------+ | VPC link | DC1 (Database) | |
| | VPC1 |-----------------| | |
| +---------------+ \ +-------------------+ |
| +---------------+ |\ |
| | VPC2 | | \ |
| +---------------+ | \ +-------------------+ |
| +---------------+ | \ VPC link | | /
| | ...... | | \__________| DC2 (Storage) |/
| +---------------+ | | |
+-------------------+ +-------------------+
Scenario: VPC to DC Connectivity
As virtualization technology becomes more and more popular, some
organizations and companies now begin to use cloud platform to
support their computer desktop, rather than using physical personal
computers and workstations. The kind of cloud platform can be a
commercial solution, e.g. Amazon's cloud platform. VPCs at cloud
service providers' DC may keep on running for a long time even if no
user is actually accessing it; but the cloud platform may bring VPCs
into power saving mode when there is little or no load in it.
The organizations and companies, e.g. a university, sometimes provide
some internal services like database which is only available to the
Cheng, et al. Expires October 4, 2015 [Page 11]
Internet-Draft Use cases for DDC Applications in SUPA April 2015
VPC but not available to users in the public network. These kind of
services can be located in an cloud operator's DC.
The VPC and the internal services sometimes are located in different
DCs, or even provided by different vendors. VPNs are configured for
the VPCs to provide connection to the internal services, and to
create and manage VPNs to internal services. The access of virtual
PCs to data resources is often controlled by underlying projects.
As shown in figure 4, service data models and policy data models can
be defined to automate the configurations of links between VPC and DC
where service is located. The data models should specify the policy
controlling authentication and authorization concerning access to
data residing in internal services. During the duration of project
X, ensure that the virtual PCs used by the project members have
secure access to the data resource Y.
7. Security Considerations
Security is a key aspect of any protocol that allows state
installation and extracting of detailed configuration states. More
investigation remains to fully define the security requirements, such
as authorization and authentication levels.
8. IANA Considerations
Not applicable.
9. Acknowledgements
The authors of this draft would like to thank the following persons
for the provided valuable feedback: Cathy Zhou, Georgios Karagiannis,
Scott O. Bradner, James Huang, Bob Natale.
10. References
10.1. Normative References
[RFC2119] Bradner, S., "Key words for use in RFCs to Indicate
Requirement Levels", BCP 14, RFC 2119, March 1997.
10.2. informative References
[I-D.karagiannis-supa-problem-statement]
Karagiannis, G., Will, W., Tsou, T., Qiong, Q., Contreras,
L., and P. Yegani, "Problem Statement for Shared Unified
Policy Automation (SUPA)", draft-karagiannis-supa-problem-
statement-02 (work in progress), October 2014.
Cheng, et al. Expires October 4, 2015 [Page 12]
Internet-Draft Use cases for DDC Applications in SUPA April 2015
[I-D.zaalouk-supa-configuration-model]
Zaalouk, A., Pentikousis, K., and W. Will, "YANG Data
Model for Configuration of Shared Unified Policy
Automation (SUPA)", draft-zaalouk-supa-configuration-
model-01 (work in progress), October 2014.
[I-D.zhou-supa-framework]
Zhou, C., Contreras, L., Qiong, Q., and P. Yegani, "The
Framework of Shared Unified Policy Automation (SUPA)",
draft-zhou-supa-framework-00 (work in progress), January
2015.
Authors' Addresses
Ying Cheng
China Unicom
P.R. China
Email: chengying10@chinaunicom.cn
JF Tremblay
Viagenie
Email: jean-francois.tremblay@viagenie.ca
Jun Bi
Tsinghua University
Bei Jing
China
Email: junbi@cernet.edu.cn
Luis M. Contreras
Telefonica I+D
Ronda de la Comunicacion, Sur-3 building, 3rd floor
Madrid 28050
Spain
Email: luismiguel.contrerasmurillo@telefonica.com
Cheng, et al. Expires October 4, 2015 [Page 13]
| PAFTECH AB 2003-2026 | 2026-04-23 15:05:31 |