Conference PaperPDF Available

Extending the UCLP software with a Dynamic Optical Multicast Service to support high performance digital media

Authors:

Abstract and Figures

Date-intensive high performance, high quality digital media traffic cannot be accommodated on traditional layer 3 networks. Alternative technologies to transmit this traffic through the network, such as optical multicast, are being investigated. A prototype of an optical multicast service was showcased during the 7<sup>th</sup> annual LambdaGrid Workshop celebrated in Prague last September. The prototype used time division multiplexing (TDM) technology as the data plane and user controlled lightpaths (UCLP) as the control/service plane. This paper describes the extensions that were done to the UCLP software to provide the dynamic optical multicast service and shows the results achieved during the Prague demonstration.
Content may be subject to copyright.
Abstract—Data-intensive high performance, high quality
digital media traffic cannot be accommodated on traditional
Layer 3 networks. Alternative technologies to transmit this
traffic through the network, such as optical multicast, are being
investigated. A prototype of an optical multicast service was
showcased during the 7
th
annual LambdaGrid Workshop
celebrated in Prague last September. The prototype used Time
Division Multiplexing (TDM) technology as the data plane and
User Controlled LightPaths (UCLP) as the control/service plane.
This paper describes the extensions that were done to the UCLP
Software to provide the Dynamic Optical Multicast Service and
shows the results achieved during the Prague demonstration.
I. MOTIVATION
Currently, a rapidly growing requirement for data
networking consists of the need to transport data intensive
digital media streams over large distances, including globally
[1]-[3]. Required transport profiles include point-to-point,
point-to-multipoint, and multipoint-to-multipoint. It has been
widely observed that this type of traffic cannot be well
supported by traditional data networking architecture and
techniques. At best, these provide for only partial, low quality
solutions that will not scale to meet the demands of new high
resolution digital media formats. For an important class of
this type of data traffic, especially for data-intensive high
performance, high quality digital media, the requirement
parameters preclude the use of common L3 techniques for
digital media. Among these parameters are those that define a
range of metrics that define high levels of quality,
performance, data volume, scale, and network resource
optimization. The requirement characteristics for these
parameters cannot be met either through common packet
routing techniques or through specialized methods for
optimizing routed performance. These streams do not display
the familiar attributes of digital media that can be observed
today in common edge devices, for example, those based on
L3 multicast techniques using standard QoS methods.
Consequently alternative techniques are being investigated.
These techniques require meeting the challenges of multiple
issues such as the dynamic allocation of major core network
resources, including lightpaths, dynamic integration of
multiple L1 and L2 paths, path duplication, edge device
addressing, path identification, path, stream, integration of
streams with different attribute characteristics (e.g., stream
size), performance monitoring and analysis, and new
mechanisms for management and operations.
To address these and related challenges, a research
consortium created an international testbed, the High
Performance Digital Media Network (HPDMnet). They are
using HPDMnet to investigate new methods for streaming
digital media, including extremely high resolution media. The
types of technologies demonstrated included new integrated
methods for discovering resources, signalling for services,
managing and controlling streams, receiving streams,
transporting streams, and duplicating streams using
dynamically allocated lightpaths. The ultimate goal of the
research consortium is to create a stable version of a global
dynamic multicast service.
UCLP, User Controlled LightPaths, is one of the
components of the HPDMnet initiative. UCLP provides a
network virtualization framework upon which communities of
users can build their own middleware or applications. The
system has been designed as a Service Oriented Architecture
(SOA) where Web Services are the basic building blocks. The
paper shows how UCLP is being extended with a new
dynamic optical multicast service. This service will provide
Extending the UCLP Software with a Dynamic
Optical Multicast Service to support High
Performance Digital Media
Eduard Grasa
1,2
, Sergi Figuerola
2
, Albert Forns
2
, Gabriel Junyent
1
and Joe Mambretti
3
1
Optical Communications Group (GCO), Signal Theory and Communications Department (TSC)
Universitat Politècnica de Catalunya (UPC), Jordi Girona 1-3, 08034 Barcelona, Spain
Tel: (+34) 93 401 7179, Fax: (+34) 93 401 7200
Email: {eduard.grasa, junyent}@tsc.upc.edu
2
Fundació i2CAT, Internet i Innovació Digital a Catalunya
Gran Capità 2-4 Edifici Nexus I 2ª planta, despatx 203, 08034 Barcelona, Spain
Tel: (+34) 93 553 2510, Fax: (+34) 93 553 2515
Email: {eduard.grasa, sergi.figuerola, albert.forns}@i2cat.net
3
International Center for Advanced Internet Research (iCAIR)
750 N. Lake Shore Drive, Suite 600, CH R317, Chicago, IL 60611 Tel: (+1) 312 503-0735
Email: j-mambretti@northwestern.edu
the user with the capability of receiving multiple high
bandwidth streams from the network and/or sending the same
stream or group of streams to one or more users in the
network using layer 1 multicast (or optical multicast)
techniques. A first instantiation of the HPDMnet was created
for the 7
th
Annual Global LambdaGrid Workshop in Prague,
and it was used to show how it was possible to stream
multiple high performance, high quality digital media streams
among multiple sites around the world from three continents
simultaneously. The optical multicast configurations were
established dynamically using a prototype of the UCLP
dynamic optical multicast service.
This paper is organised as follows. Section II provides an
overview of UCLP and its service oriented architecture.
Section III details the changes that have been done to the core
services of UCLP to support multicast, and explains the
design of the Global Dynamic Multicast Service. Section IV
shows the optical multicast demonstration carried out in
Prague during the 7
th
Annual Global LambdaGrid Workshop
last September. Finally Section V summarizes the conclusions
and gives an overview of the future directions of the Optical
Multicast and the UCLP projects.
II. UCLP
A. Introduction
During recent years there has been a trend for organizations
such as universities, hospitals, schools and big corporations to
buy their own fibre and point-to-point links from various
suppliers [4]. These organisations would like to manage these
resources as part of their own network domain, and, at the
same time offer some services to their end users. Another new
phenomenon is that many large remote sensor- and
instrumentation-related projects are appearing that generate
large data flows that need to be delivered to computing centres
that can be far away. Consequently, an emerging new type of
users is appearing that want to remotely control these
instruments and modify the topology of the network
connecting these data sources and sinks from time to time.
Nowadays, the configuration of the scenario described above
would be set up statically or it would have been pre-
configured by an operations network engineer.
The UCLP software provides an alternate network
provisioning process which allows users to control their own
packet- or switched-based network architecture including
topology, routing, virtual routers, switches, virtual machines
and protocols. The UCLP concept consists of many separate,
concurrent and independently managed Articulated Private
Networks (APNs) operating on top of one or more network
substrates across different ownership domains. An APN can
be considered as a physically isolated or “underlay” network
where a user can create a customized multi-domain network
topology by binding together layer 1 through 3 network links,
computers, time slices and virtual or real routing and/or
switching nodes. This UCLP capability is realized by
representing all network elements, devices and links as web
services, and by using web services workflow to allow users
to bind together their various web services to create a long-
lived APN.
B. Software Architecture
The UCLP service-oriented architecture is depicted in Fig
1. Each box represents a different service. The Resource
Management Services or Network Element Web Services is a
group of services that manage and control the resources on a
physical device; each single web service dealing with a
different technology: the Cross Connect Web Service (XC-
WS) controls devices that perform cross connections (TDM,
wavelength, fibre); the GMPLS Web Service (GMPLS-WS)
deals with GMPLS networks; the 802.1q Web Service
(802.1q-WS) manages VLAN-enabled equipment; the MPLS
Web Service (MPLS-WS) controls MPLS networks and the
Logical Router Web Service (LR-WS) controls logical as well
as physical routing platforms. Resource Virtualisation
Services provide a layer of virtualisation so that the
technology of the physical devices is abstracted but none of
the features that these devices offer are lost. Finally, higher
level services or applications exploit the virtualisation
capability just described in order to build complete end user
solutions or other services without having to deal with the
underlying network complexities.
Fig. 1. UCLP Service Oriented Architecture. All the UCLP Components are
designed as different web services. By selecting one or more of these
components and stitching them together, different network resource
management solutions can be created.
The first higher level service designed and implemented is
the APN Scenarios WS. APN Scenarios are “snapshots” of the
network, i.e. a given configuration of the network elements
and links that are part of the APN. The APN Scenarios WS
allows a user to design one or more scenarios and to put them
in place in a single transaction. For instance, the user makes a
request to the APN Scenarios service and tells it to put
scenario number one in place. The APN Scenarios service will
issue all the required calls to the other WSs to ensure that all
the configurations defined in the scenario one are executed on
the network elements. The single transaction features allows
the APN Scenarios service to roll back when any of the
predefined network element configurations cannot be applied.
Any of the service groups presented in the architecture can
be extended to support new technologies or to provide new
capabilities in the virtualisation layer. A more in-depth
explanation of the architecture can be found in [5], and an
example of a hypothetical deployment of the UCLP software
over the paneuropean network is given in [6].
III. EXTENDING
UCLP TO SUPPORT OPTICAL
MULTICAST
The first research and implementation prototypes of optical
multicast have been targeted to the Time Division
Multiplexing technology (SONET and SDH), particularly to
the Nortel Optical Multiservice Edge 6500 (Nortel OME
6500) and Nortel Optical Cross Connect HDXc (Nortel
HDXc) platforms. Both platforms provide the “Drop and
Continue” functionality that allows a data stream in an input
port to be replicated up to N-1 times and forwarded to N
output ports (the N-1 copies plus the original stream). The
OME platform allows you to create up to 3 copies (so the split
ratio can be up to 1:4), and the HDXc platform only allows
you to create 1 copy (split ratio of 1:2). Data replication is
performed on real time, so the end-to-end delay and jitter
experienced by the data stream are not affected.
UCLP already supported part of the command set of the
Nortel OME 6500 and the Nortel HDXc platforms, but the
core UCLP services and the Graphical User Interface (the
GUI, also called Resource Management Centre) could only
handle point to point connections; therefore an extension to
support multicast connections, explained on subsections A and
B, has been implemented.
A. Core services
TDM technology is handled by the Cross-Connect Web
Service (XC-WS in Fig. 1). The web service interface of the
XC-WS has been extended with the ability to request for one-
to-many, many-to-one and many-to-many connections (only
one-to-many and many-to-one are used for the multicast
framework). The signalling of a multicast connection both in
the Nortel OME 6500 and the Nortel HDXc platforms
requires the exchange of multiple configuration messages with
the physical devices, so a transactional type of behaviour
when configuring the devices must be ensured (i.e. either all
the required configurations are successful or, if one fails, the
previously applied configurations are rolled back and the
device is left on the initial stage).
The commands required to create a 1:4 multicast
connection in the Nortel OME 6500 are the following (using
the TL1, Transaction Language 1, remote interface):
Create 4 point to point connections between the input
port and each one of the output ports, using the ENT-
CRS TL1 command.
The TL1 command flow to create a 1:2 multicast
connection in the Nortel HDXc platform is the following:
Create a point to point connection between the input port
and one of the output ports, using the ENT-CRS TL1
command.
Create the second connection, between the input port and
the other output port, using the ENT-ROLL TL1
command.
Resource Virtualization Services (see Fig. 1) are the
responsible for storing the connection data structure. Prior to
the multicast extension, the data structure was an ordered list
of all the resources that were part of the connection. This
structure has been replaced by a tree data structure. When a
multicast connection has to be undone, the Resource
Virtualization Services parse the tree, retrieve what resources
are connected in each network element and issue the required
calls to the Resource Management Services to undo the
connections. A similar procedure is carried out when creating
the multicast connections.
Fig. 2. Example of the solution tree computed by the algorithm that finds the
multicast tree with the minimum number of resources. In the example there is
one source, s, and two destinations, t
1
and t
2
.
B. Resource Management Centre
UCLP Resources (links and interfaces) are managed
through the Resource Management Centre (RMC), the
Graphical User Interface. The RMC allows users to see and
manipulate the resources they can access. For instance, they
can trade the resources (i.e. lease resources to other users or
acquire more resources from other users), they can partition
and bond them and they can connect them in any (possible)
way they want. To help the user create connections with their
resources the RMC is equipped with two route selection
functions: the first one, based on Dijkstra’s shortest path
algorithm, selects the minimum number of resources to create
a connection; the second one, called “All Routes”, calculates
all the possible routes between a source and a destination pair
(given a maximum number of hops specified by the user), and
lets the user choose the route he wants the connection to
follow. Both algorithms only support point to point
connections; therefore two new algorithms have been
implemented to handle point to multipoint connections.
The first algorithm is the equivalent to Dijkstra’s shortest
path algorithm, is used to select the minimum number of
resources that will create a valid multicast tree given a source
interface, a list of destination interfaces and the desired
bandwidth. The shortest path algorithm creates a graph with
all the nodes and links and calculates all the possible routes
between the source node, s, and each one of the target nodes,
t
i
. So for each pair {s,t
i
} a list with all the possible routes
between s and t
i
is generated. The complete list of routes
between all source and destination pairs is used to create a
solution tree using a backtracking algorithm. The solution tree
has the source as a root and as many levels as targets exist in
the multicast connection. The root element has n children
leaves, being n the number of possible routes between s and t
1.
Each one of the n
1
leaves of the first level of the tree stores a
list of the resources used in that particular route between
s and
t
1
. Each leave of the first level of the solution tree has n
2
children leaves, being n
2
the number of routes between s and
t
2.
Each one of the n
2
* n
1
leaves of the second level stores the
resources resulting from the union of two sets: the resources
used in the route n
1
between s and t
1
and the resources used in
the route n
2
between s and t
2.
Further levels of the solution tree
(if they exist) are computed the same way. The optimal
solution is the leave of the last level that has fewer resources
stored. An example of the solution tree of a multicast
connection with one source and two destinations is given in
Fig. 2. As the number of leaves in each level grows
exponentially, it is very inefficient to compute the complete
solution tree. This is why a backtracking algorithm that uses
the “Branch and Cut” approach is used to generate the
solution tree. The algorithm will only keep developing a
branch if the number of resources stored in the current leave
of the branch is less than the number of resources of a
temporary solution (the temporary solution is the minimum
number of resources of all the already computed multicast
trees). If this condition is not fulfilled, the algorithm will cut
the branch and will continue developing a new one, until there
are no more branches to develop.
The second algorithm is very similar to the first one, the
only difference being that the solution must compute all the
possible multicast trees. Therefore, the same solution tree as in
the first algorithm is created, but this time the backtracking
algorithm does not cut any branch. Because the number of
possible multicast trees can be overwhelming, the user
interface only displays a limited number of solutions
(configurable by the user).
C. GDOM Service
Once the core services of UCLP support multicast
connections, a higher level service can be created on top of
them to allow the user to interact with a higher level
abstraction. This way the user can use the multicast service in
a more intuitive an simple fashion, because some complexities
associated with the provisioning of the multicast service, such
as route selection, the modification of existing multicast
connections without tearing them down, and others will
remain hidden from the user interface.
Logical rings, depicted in Fig. 3, are the abstraction chosen
for the Global Dynamic Optical Multicast (GDOM) Service.
Fig. 3a shows a logical ring populated with several streams.
When a new user site wants to become a member of the ring,
he can choose what streams he wants to download from the
ring to the site and what streams he wants to upload from the
site to the ring, as seen in Fig 3b. It is interesting to note that if
a site downloads a stream from the ring, the stream is not
eliminated from the ring and other sites can still download it.
This feature is possible due to the fact that the ring is multicast
enabled and many to many communication between user sites
is supported. Logical rings are a good abstraction because they
let the user express their intentions in terms of what he really
cares: what are the streams that he wants to get from the
network and what are the streams that he wants to send to the
network. All the complex processes that must be carried out to
fulfil the user’s requests are hidden from them and are
performed by the internal logics of the GDOM service.
Fig. 3a. A logical ring. All the sites connected to the ring can send streams to
the ring and can receive streams from the ring. The same stream can be
dropped in several sides due to the ring multicast capabilities.
Fig. 3b. The ring interface. Each site can decide which streams it wants to get
from the ring and which it wants to upload to the ring.
The GDOM service is currently in its design phase, but its
functionality is already planned. Each GDOM service will be
able to manage one or more logical links. In a first stage the
different logical rings managed by the GDOM service will be
isolated from each other, but inter-ring communication will be
handled once the basic functionality is implemented and
tested. For each logical link, the GDOM service will have to
dynamically create and modify multicast connections between
the various sites participating in the ring. It is important to
note that the service must have the ability to modify existing
unicast or multicast connections without tearing them down.
For instance, imagine that, in Fig 2a. “Site B” and “Site C” are
already receiving a stream contributed by “Site A”. Then “Site
H” arrives and makes a request to receive the “Site A” stream.
The GDOM service will have to add another receiver to the
already existing multicast connection between “Site A”, “Site
B” and “Site C” without affecting “Site B” and “Site C”. This
is achieved by adding another branch to the existing multicast
tree between “Site A”, “Site B” and “Site C”. The algorithms
required to optimally choose the resources of the new branch
so that each multicast tree uses the minimum number of
available resources will be an adaptation of the algorithms
designed for the Resource Management Centre, explained in
the previous section. Once the resources of the new branch of
the tree have been selected, the GDOM Service will issue
requests to the UCLP Core services so that the required
configuration changes are applied to the network elements.
IV. D
EMONSTRATION AT THE 7TH ANNUAL GLOBAL
LAMBDAGRID WORKSHOP
GLIF [7], the Global Lambda Integrated Facility, is an
international virtual organization that promotes the paradigm
of lambda networking. GLIF provides lambdas internationally
as an integrated facility to support data-intensive scientific
research, and supports middleware development for lambda
networking. A subset of the GLIF resources was allocated to
create the HPDMnet instantiation used to showcase the
capabilities of next generation digital media services using
lightpaths, during the 7th Annual Global LambdaGrid
Workshop held in Prague.
During the demonstration, five sites from different places in
the world streamed high performance digital video content to
each other and to the demo venue, where large screens
showed diverse media content including cultural events and
historic architecture (Barcelona-i2CAT), a live MusicGrid
lesson and Canadian landscapes (Ottawa-CRC and Ottawa-
Nortel), nanotech virtual instrumentation (Chicago-iCAIR),
and experimental images (Amsterdam-SARA). Each stream
consumed a bandwidth of around 750 Mbps, a requirement
that can be esasily fulfilled using optical networks but that is
very hard to achieve using a shared IP network.
Time Division Multiplexing (SONET and SDH) was the
technology of choice for the demo. The network elements
responsible for creating the copies of the streams were the
Nortel OME 6500 Platform and the Nortel HDXc Platform.
Fig. 4 illustrates the demo scenario. Each colored line
represents a different stream, and each dashed square a
different site. All the multicast connections were defined prior
to the demo using the UCLP Resource Management Center,
and were stored as an APN Scenario. During the demo, the
Fig 4. Layer 1 scenario of the dynamic optical multicast demo carried out at Prague. Each colored line represents a stream originated at a different site.The
Nortel OME 6500 and the Nortel HDXc platforms were used to replicate the streams and create the different multicast distribution trees.
APN Scenario was instantiated and UCLP configured all the
required cross-connections on the network elements to create
the multicast trees. After a certain period of time, the APN
Scenario was torn down and the UCLP software undid all the
cross-connections. This sequence of events was performed
several times during the demonstration.
The demonstration in Prague was designed to show a work-
in-progress, not a completed service or product. Further
research will be conducted by the consortium on these
methods and technologies. Also, additional demonstrations are
planned at other events scheduled for later this year. The next
one will take place at ArtFutura 2007 [8], in the Mercat de les
Flors in Barcelona, next 25-28 October 2007.
V. C
ONCLUSION AND FUTURE WORK
Even though the technologies explained in this paper are
just a first prototype of the service under investigation, the
feedback received after the demonstration showed that there is
a lot of interest in setting up a service that allows high
performance media streams to be sent to one or more locations
through the network. Therefore the consortium will continue
the research towards a stable and “production-grade” version
of the service. Future research directions can be split in two
main topics: data plane technologies and control/service plane
software. The former will try to provide the optical multicast
service using the latest advances in Ethernet technologies as
an alternative to TDM, while the latter will deal with
providing a higher level of service abstraction as a basis to
interact with the user so that the service can be setup in a more
intuitive and easier fashion (for instance, using the “logical
ring” model explained in this paper).
Regarding UCLP, it can be concluded that the software has
proven to be extensible and that it can fulfil the control
requirements for the optical multicast service. Future work in
the UCLP initiative includes creating a commercial,
production-grade version of the software to control optical
resources called “Argia” (optical in a broad sense of the word,
meaning fibre, lambda, TDM and carrier Ethernet) resources.
UCLP research efforts will also continue, focusing in
extending the software to new technologies (vlan-based
equipment, logical routers) and designing innovative services
for the user, such as the GDOM service.
A
CKNOWLEDGMENT
The authors would like to thank Michel Savoie, Scott
Campbell and Hanxi Zhang from the Communications
Research Centre (Canada), Mathieu Lemay from Inocybe
Technologies Inc. (Canada), Albert López from UPC (Spain),
Eric Bernier from Nortel Networks (Canada), the NOC
engineers from Starlight, Netherlight and Catlight and Hervé
Guy and Thomas Tam from CANARIE Inc. (Canada) for their
contributions to this work.
R
EFERENCES
[1] J. Leigh et al, “The Global Lambda Visualization Facility: An
international ultra-high-definition wide-area visualization collaborator”,
Future Generation Computer Systems, volume 22, issue 8, pp. 964-971.
October 2006.
[2] T. Shimizu et al. “International real-time streaming of 4K digital
cinema”, Future Generation Computer Systems, volume 22, issue 8, pp.
929-939, October 2006.
[3] CineGrid. [Online]. Available: http://www.cinegrid.org/
[4] Dark fibre case studies provided by NetFiber [Online]. Available at:
http://www.nefiber.com/case-studies
[5] J. Wu, M. Savoie, S. Campbell, H. Zhang, and B. St. Arnaud, "Layer 1
Virtual Private Network Management by Users", IEEE Communications
Magazine, Feature Topic on Advances in Control and Management of
Connection Oriented Networks, pp. 86-93, December 2006.
[6] E. Grasa, S. Figuerola, A. López, G. Junyent, M. Savoie, B. St. Arnaud,
M. Lemay, “Articulated Private Networks in UCLP”, TERENA
Networking Conference, May 2007 [Online]. Available:
http://tnc2007.terena.org/programme/presentations/
[7] The Global Lambda Integrated Facility [Online]. Available at:
http://www.glif.is/
[8] ArtFutura 2007 [Online]. Available at:
http://www.artfutura.org/english/index.php
Article
Cloud computing services have recently become a ubiquitous service delivery model, covering a wide range of applications from personal file sharing to being an enterprise data warehouse. Building green data center networks providing cloud computing services is an emerging trend in the Information and Communication Technology (ICT) industry, because of Global Warming and the potential GHG emissions resulting from cloud services. As one of the first worldwide initiatives provisioning ICT services entirely based on renewable energy such as solar, wind and hydroelectricity across Canada and around the world, the GreenStar Network (GSN) was developed to dynamically transport user services to be processed in data centers built in proximity to green energy sources, reducing Greenhouse Gas (GHG) emissions of ICT equipments. Regarding the current approach, which focuses mainly in reducing energy consumption at the micro-level through energy efficiency improvements, the overall energy consumption will eventually increase due to the growing demand from new services and users, resulting in an increase in GHG emissions. Based on the cooperation between Mantychore FP7 and the GSN, our approach is, therefore, much broader and more appropriate because it focuses on GHG emission reductions at the macro-level. This article presents some outcomes of our implementation of such a network model, which spans multiple green nodes in Canada, Europe and the USA. The network provides cloud computing services based on dynamic provision of network slices through relocation of virtual data centers.
Conference Paper
Most grid applications require the processing of large amounts of data stored at different locations across the network which makes optical grid infrastructures optimal for such applications. The increase in intensity of data- and communications of these applications calls for new mechanisms and theories on how to optimally allocate optical grid resources. Typical Divisible Load Theory (DLT) makes optimal allocation of computational resources. In this paper, we introduce Network Aware Divisible Load Algorithm (NADLA) that extends DLT to optimally allocate both the computational and networking resources of an optical grid. We assume a data- and communications-intensive application where data is stored at different sites across the optical grid and can be divided into independent subsets to be processed in parallel at different sites. The algorithm defines an optimal data transfer schedule that defines when to transfer what data subsets to which sites across the optical grid in order to minimize the overall application completion time. It consists of two phases. In the first phase, NADLA provide load distribution that minimizes application completion time taking into account network connectivity and computational and networking resources availability. Site connectivity is estimated by considering the bandwidth and the expected free time of all the links connecting this site to other sites. In the second phase, a simple greedy algorithm is used to allocate computational and networking resources. Extensive Simulations are conducted to examine the performance of the proposed algorithm for different application types and sizes, and different optical grid topologies. Simulation results show the advantages of the proposed algorithm over the traditional DLT approach for the category of applications considered.
Article
Full-text available
The research outlined in this paper marks an initial global cooperative effort between visualization and collaboration researchers to build a persistent virtual visualization facility linked by ultra-high-speed optical networks. The goal is to enable the comprehensive and synergistic research and development of the necessary hardware, software and interaction techniques to realize the next generation of end-user tools for scientists to collaborate on the global Lambda Grid. This paper outlines some of the visualization research projects that were demonstrated at the iGrid 2005 workshop in San Diego, California.
Article
Full-text available
Purpose – The purpose of this paper is to illustrate the advantages of using the UCLP software for network operators, advanced and regular end users in the research networking community. Design/methodology/approach – This paper provides an example of the deployment of UCLP in the GÉANT2/National Research and Education Networks (NRENs) scenario, and compares how network operators, advanced users and regular end users would do their work, with and without UCLP. Findings – The paper provides high-level technical information about UCLP as well as depicting the drivers for its use in the European research networking community. Research limitations/implications – This paper does not explain the details of the deployment of the software in the GÉANT2/National Research and Education Networks (NRENs) scenario, it just explains the benefits that the deployment of the software would provide. If the deployment was to be done today, some improvements to UCLP should be done, as well as support for more equipment vendors should be added. Practical implications – UCLP could provide more flexibility to the e-science community if it was deployed over the European research networking infrastructure, because it would isolate network users from each other while providing them an unprecedented degree of control over the network. Originality/value – Nowadays, several control/management solutions for networks exist, but none that is capable of partitioning a physical network into slices and handoff its management to the users, like UCLP does. This is the first UCLP paper that studies a hypothetical deployment of UCLP in the European research networking scenario, and evaluates the drivers and implications of such a deployment.
Article
Full-text available
The layer 1 virtual private network (LlVPN) technology supports multiple user networks over a common carrier transport network. Emerging L1VPN services allow: L1VPNs to be built over multiple carrier networks; L1VPNs to lease or trade resources with each other; and users to reconfigure an L1VPN topology, and add or remove bandwidth. The trend is to offer increased flexibility and provide management functions as close to users as possible, while maintaining proper resource access right control. In this article two aspects of the L1VPN service and management architectures are discussed: management of carrier network partitions for L1VPNs, and L1VPN management by users. We present the carrier network partitioning at the network element (NE) and L1VPN levels. As an example, a transaction language one (TL1) proxy is developed to achieve carrier network partitioning at the NE level. The TL1 proxy is implemented without any modifications to the existing NE management system. On top of the TL1 proxy, a Web services (WS)-based L1VPN management tool is implemented. Carriers use the tool to partition resources at the L1VPN level by assigning resources, together with the WS-based management services for the resources, to L1VPNs. L1VPN administrators use the tool to receive resource partitions from multiple carriers and partner L1VPNs. Further resource partitioning or regrouping can be conducted on the received resources, and leasing or trading resources with partner LlVPNs is supported. These services offer a potential business model for a physical network broker. After the L1VPN administrators compose the use scenarios of resources, and make the use scenarios available to the L1VPN end users as WS, the end users reconfigure the L1VPN without intervention from the administrator. The tool accomplishes LlVPN management by users
Article
This paper describes the world’s first real-time, international transmission of 4K digital cinema and 4K Super High Definition (SHD) digital video at iGrid 2005, hosted at the California Institute of Telecommunications and Information Technology (Calit2) at the University of California, San Diego. Nearly six hours of live and pre-recorded 4K motion picture and audio content was streamed to iGrid in San Diego from the Research Institute for Digital Media and Content (DMC) at Keio University in Tokyo.To implement this demonstration, several new technologies were introduced, including a prototype high-performance 4K compressed multicasting system called “JPEG 2000 Flexcast”, and “Soundscape”, a practical scheme for synchronizing audio and video transmitted from different locations over IP networks.These iGrid 2005 demonstrations proved that it is now feasible to implement networked professional audio/video applications–production, post-production and distribution–even at 4K quality over IP networks up to 15,000 km long. The demonstrations also showed the new 4K motion picture technology being introduced for digital cinema can be usefully applied to other network applications such as remote telepresence, distance learning and scientific visualization.
Feature Topic on Advances in Control and Management of Connection Oriented Networks
  • J Wu
  • M Savoie
  • S Campbell
  • H Zhang
  • B St
  • Arnaud
J. Wu, M. Savoie, S. Campbell, H. Zhang, and B. St. Arnaud, "Layer 1 Virtual Private Network Management by Users", IEEE Communications Magazine, Feature Topic on Advances in Control and Management of Connection Oriented Networks, pp. 86-93, December 2006.
Articulated Private Networks in UCLP
  • E Grasa
  • S Figuerola
  • A López
  • G Junyent
  • M Savoie
  • B St
  • M Arnaud
  • Lemay
E. Grasa, S. Figuerola, A. López, G. Junyent, M. Savoie, B. St. Arnaud, M. Lemay, "Articulated Private Networks in UCLP", TERENA Networking Conference, May 2007 [Online].