|
Cisco uses the term GLN Cradle to refer to the hardware,
software, and applications that
run the Global Learning Network and that currently reside
at the Cable and Wireless Internet
Services data center in San Jose, California. These applications
are housed in the cradle:
 Content Distribution System
Virtual Classroom
CNAMS
CLI Virtuoso
Managed Services
The San Jose data center (and others as GLN grows) has redundant
connections, network
devices, and servers with no single point of failure. There
is a T3 connection through a Cisco
7200 Series router. After passing through the Cisco PIX Firewall,
LocalDirector load balances
server requests across a server farm to connect to multiple
Cisco Catalyst 6509 Series
switches. In the event of a failure at any point on the network,
backups are in place to
immediately take over with minimal or no interruption in service.
The LAN components within the data center also have full
redundancy, protecting the network
against WAN outages and the failure of a LAN device. The architecture
maximizes network
uptime, while minimizing the potential for network failures.
Equipment supported in or by the GLN Cradle:
LAN
The Local Area Network (LAN) at each Networking Academy supports
the local Web server, the Content Engine if implemented, and
the connection to the Internet. Academies are urged to utilize
the GLN Services Manager to analyze their network topology
before a Content Engine is implemented.
Web Server
All Academies are expected to maintain a local Web server.
Static curriculum may be downloaded to the local Web server
and then made available to students. A local Web server allows
students to access curriculum faster by not having to constantly
go out over the Internet to view content. This also keeps
network traffic at a manageable level and provides a cost
avoidance opportunity by minimizing dynamic content delivery
charges.
Content Distribution Manager (CDM)
The GLN architecture was designed to support very large files
containing streaming video, complex graphics, and audio. The
Cisco Content Delivery Network System (CDN) is an integrated
system for intelligently pre-positioning content at the network
edge for faster, high-quality delivery of learning curriculum.
The GLN CDN consists of Cisco Content Distribution Managers
(CDM 4650s) at the Cable and Wireless Internet Services data
center in San Jose, California; Cisco Content Engines (CE-507s)
that store media-rich content at local Networking Academy
sites; and the software that controls the flow of content.
Currently, the network includes one active CDM and a second
hot standby system. The primary CDM coordinates the activities
of the CDN including bandwidth allocation, content replication,
and device management.
The CDM also monitors CE connectivity to ensure that Academy
CEs are online and able to receive content at their scheduled
times. Academies simply indicate in their site surveys the
times they want to receive content and the bandwidth that
is available for the CDM to use.
The CDM redirects user requests to the local CE for delivery
of the Networking Academy CCNA curriculum 2.1.4E, an enriched
media curriculum that features streaming video, Flash animations,
zoom photography, and topic drill down capability through
More Info buttons.
Content Engine
A Cisco Content Engine delivers streaming media to students.
The content engine functions much like a file server. Its
form-factor is small since it is a 1-U rack-mounted server
and looks somewhat like a small Cisco switch. The optional
content engine typically resides on the Networking Academy's
LAN. Using the CDM control panel, a CE can subscribe to multiple
channels offered by the CDM, and each channel contains unique
content.
To save time and bandwidth, individual CEs are updated by
the CDM with data from only those channels in which the CE
is subscribed. Logically, these CE updates only occur when
the subscribed CDM data has been updated. If a change is made
to a channel that a Networking Academy does not subscribe
to, the Academy is not updated or affected by the change.
In a scalable architecture, the goal is to keep the updates
limited to only what is necessary. By creating channels for
content and determining what has actually changed within a
channel, updates are very specific and focused. The CEs can
stream content at about 10/100 megabytes per second. One CE
can handle up to 30 students simultaneously viewing a content
item. Therefore, 40 students viewing the same movie would
require two CEs.
When a student is viewing content via the Web server and
clicks the "More Info" option button, a request
is made to the CLI Virtuoso Delivery system to display a list
of all available media files for that topic. The Content Delivery
Manager makes decisions about where the student is, what technology
is available, and the closest point content can be delivered
to, typically to the local CE. In the Academy implementation,
there are various scenarios where CEs can be used in either
a single classroom or even a single school, depending on the
number of students participating and the number of students
who are using the CE at the same time. Initially, channels
will be grouped according to content and language with English
being the only "More Info" language currently supported.
Network Switch
In the GLN architecture, switching services are provided by
Cisco Catalyst 6509 switches implemented in a redundant configuration
in the GLN Cradle. The Catalyst 6500 series is designed to
deliver highly available, secure, converged network services
for both Enterprise and Service Provider networks. This architecture
addresses the increased requirements for scalability, high-availability,
rich services, and multi-layer switching in backbone, distribution,
and wiring closet topologies as well as data center environments.
The Catalyst 6500 series delivers exceptional scalability
and performance, supporting a wide range of interface densities,
performance, and integration of powerful services modules.
Routers
Routers are the traffic cops of the Internet highway, directing
packets to their proper destination, one hop at a time. Routers
make use of routing protocols that are configured to establish
the metric for directing traffic to other networks. Routing
protocols establish where to route traffic when there are
multiple paths to a destination network and where to route
traffic when routing activity is so busy that traffic would
be better directed to another router. In the Networking Academy
implementation, incoming traffic may be checked against access
lists (protocols and addresses). Non-approved traffic is blocked.
Approved traffic is routed through another circuit to the
firewall and on to the Web servers. The routers used by the
Networking Academy are the Cisco 7200 Series, model 7206 VXR
routers. There are at least two routers in place in each data
center connected together with Hot Stand-by Routing Protocol
(HSRP)-Cisco's disaster recovery routing protocol.
Internet/ISP
The Internet Service Provider (ISP) is not part of the GLN
architecture but is a necessary element for GLN functionality.
The ISP provides Internet connectivity for a Networking Academy.
Typically, an ISP provides the circuits and data centers,
although in some large installations these reside at the Academy.
For implementation of the Cisco Networking Academy Program,
Cable and Wireless Internet Services provides all the circuits
among their data centers; however, numerous ISPs are similarly
equipped to provide data center connectivity. Each Academy
connects to the Internet through an ISP of their choosing.
Academies must maintain a minimum 56K dedicated circuit. Soon
that requirement will be increased to128K.
DistributedDirector
The DistributedDirector, provided by the hosting data center,
routes a user to the most appropriate data center, based on
logical proximity and the response times of the data center.
In the Networking Academy implementation, Cable and Wireless
Internet Services uses a Cisco DistributedDirector in the
San Jose data center. Additional DistributedDirectors will
be placed in different data center locations around the world
as the network demand increases. DistributedDirector looks
at the availability and timing of the Web servers and routes
traffic to the closest Web server, based on pre-configured
decisions.
LocalDirector
Load balancing devices, like the Cisco LocalDirector, allow
for Web server redundancy. LocalDirectors are configured to
know how many Web servers there are so they can balance the
loads among busy and less active servers by checking the frequency
of their responses. If a Web server goes down, the LocalDirector
routes traffic to other Web servers that are fully functional.
The LocalDirector frequently checks the status of the Web
servers and will stop directing traffic to a non-responding
Web server without any manual intervention.
The Networking Academy implementation uses two Cisco LocalDirectors
per data center, one primary and one secondary. A failover
cable set up between the two LocalDirectors facilitates communication
between them. If one of the LocalDirectors fails to respond,
the active LocalDirector takes over.
Firewalls
Firewalls protect all the equipment behind them by restricting
access to certain devices and certain traffic. All communications
must comply with the rules in the firewall. Firewalls use
access lists to allow only certain types of traffic to go
to specified addresses. Another feature used by firewalls
is Network Address Translation (NAT), which translates the
expected address to another address that the equipment actually
responds to.
In the Academy implementation, the Web servers are located
behind two PIX Firewalls at the data centers. The two firewalls
employ a redundant, failover architecture and are in constant
communication with one another. If communications are disrupted,
the firewall that is operational automatically takes all the
traffic.
Delivery and CNAMS Web Servers
The Delivery and Cisco Networking Academy Management System
(CNAMS) Web servers are the Web interfaces between the users
at the local Academies and the applications, content, and
services they need.
Content Servers
Students can access curriculum from within their LAN if the
curriculum has been previously downloaded to the local Web
server at the Academy. Once the curriculum has been downloaded,
service provider content delivery charges are avoided for
subsequent access to the curriculum.
When the curriculum is not on the local Web server, student
requests for curriculum are directed to the content server
at the data center with the nearest logical proximity. The
content server behaves like a middleman between the CLI Virtuoso
Delivery database and the Web servers that store cached content
for downloading by users. Data center Web servers interface
with the end user through HTTP requests and also communicate
with the content server. Content servers are used to process
user requests and to make decisions on where to obtain the
requested content. Assessments are never cached and are accessed
through the CLI Virtuoso Delivery database.
CLI Virtuoso Delivery Database
The curriculum database server is where all curriculum and
assessment items reside. The CLI Virtuoso Delivery database
is responsible for delivering curriculum and generating personalized
feedback. CLI Virtuoso Delivery facilitates dynamic Web page
creation and delivery based on the individual learner's grasp
of course materials. Web pages are created automatically by
the system as the learner completes the class assessments.
CLI Virtuoso Delivery database is also responsible for displaying
a list of optional "More Info" media files for a
specific Target Indicator when requested by the user. When
a media file is selected, the request is passed to the CDM
and the CDM re-directs the request to retrieve the file from
the available Academy content engine.
CNAMS Database
The Cisco Networking Academy Management System (CNAMS) database
is the management component of the Cisco Networking Academy
Program. The CNAMS database handles authentication and accessibility
to different types of curriculum.
|
