USF IT Video Part 3
Summary
TLDRThis video script delves into the University of South Florida's (USF) network infrastructure, explaining how desktops connect to the network and communicate with other devices. It highlights the role of core routers in providing internet connectivity through various ISPs and the use of BGP for efficient routing. The script also discusses the importance of data centers for hosting business data and applications, the redundancy measures in place, and the network's monitoring and management practices. It emphasizes the need for enterprise-class equipment to ensure reliability and performance in a large-scale network environment.
Takeaways
- 🖥️ Desktops on the USF network need to communicate with other devices like servers and clients to be useful.
- 🌐 The USF network uses a pair of Wang routers and a Cisco ASR 9006 to provide wide area connectivity and internet service through various ISPs.
- 🔌 USF obtains internet connectivity through connections like Cogent, Time Warner Telecom, and Florida Lambda Rail, delivered over single-mode fiber as Ethernet connections.
- 🛣️ Border Gateway Protocol (BGP) is used by LAN routers to determine the best ISP for routing traffic to its destination based on approximately 450,000 advertised routes.
- 🔄 When a remote server responds, BGP routing tables guide the traffic back to USF, ensuring reliable communication.
- 🏢 USF's data centers host business data and applications critical to the university's operations, with redundancy in place to ensure continuous service.
- 🔌 The Tampa data center is directly connected to the campus core, with servers connecting at speeds of 1 Gbps and 10 Gbps to access layer switches.
- 🛡️ Network gear in data centers is enterprise class, designed for reliability, security, and remote management, unlike SOHO class gear suitable for small offices or homes.
- 🔍 Network engineering group monitors network devices every three minutes for availability, health, and performance, addressing any issues promptly.
- 📈 Extensive monitoring of network statistics aids in capacity planning, troubleshooting, and security, covering everything from traffic levels to telephony system call volumes.
- 👥 A small team of experts manages and operates the nearly 10,000 devices of the USF network, highlighting the importance of enterprise-class equipment for large-scale network management.
Q & A
What is the primary purpose of networking?
-The primary purpose of networking is to enable communication between different devices, such as servers and clients, to facilitate data sharing and resource accessibility.
How does traffic from a USF desktop reach the Internet?
-Traffic from a USF desktop reaches the Internet by being routed through the campus backbone to WAN routers, which then determine the best ISP to use for reaching the destination IP address based on BGP routing tables.
What are the two types of routers mentioned in the script that provide wide area connectivity for USF?
-The two types of routers mentioned are a Wang router and a Cisco ASR 9006, which work together to provide wide area connectivity to USF's Internet service providers.
How does USF obtain Internet connectivity?
-USF obtains Internet connectivity through connections with Cogent, Time Warner Telecom, and Florida Lambda Rail, delivered as Ethernet connections over single-mode fiber.
What protocol do the LAN routers at USF run to make routing decisions?
-The LAN routers at USF run BGP (Border Gateway Protocol) to make routing decisions for traffic destined for the Internet.
What is the significance of having multiple data centers for USF?
-Having multiple data centers, including a primary, a secondary, and a disaster recovery center, ensures continuous service and data redundancy in case of failures or disasters.
How are the servers within USF's data centers connected?
-Servers within USF's data centers are connected at 1 gigabit per second and 10 gigabit per second speeds to access layer switches, which then connect to the core network equipment.
What measures are taken to ensure high availability and redundancy in USF's network?
-USF ensures high availability and redundancy by having dual power supplies, network connections to two different switches for critical servers, and uplinks to different Cisco 6500s to maintain connectivity in case of a switch failure.
How frequently does the network engineering group monitor the network devices at USF?
-The network engineering group monitors the network devices at USF every three minutes, checking for availability, health, and various performance parameters.
What types of statistics does USF monitor to maintain and troubleshoot its network?
-USF monitors statistics such as input and output traffic levels, concurrent users, throughput on wireless access points, network performance across backbone and inter-campus LAN links, and call volumes on telephony systems.
What is the difference between enterprise class and SOHO class network equipment as mentioned in the script?
-Enterprise class network equipment is designed for large networks, offering reliability, security, remote management, and high performance, whereas SOHO (Small Office/Home Office) class equipment is designed for smaller environments and prioritizes ease of use and affordability.
How many engineers are responsible for designing and supporting all of USF's networks and services?
-A small group of five engineers is responsible for designing and supporting all of USF's networks and services.
Outlines
🌐 Network Communication and Infrastructure at USF
This paragraph explains the importance of network connectivity beyond just having a desktop connected to the University of South Florida (USF) network. It details how traffic is routed to the Internet once it reaches the USF campus core through a pair of Wang routers and a Cisco ASR 9006, highlighting the use of BGP for determining the best ISP path to a destination IP. The paragraph also discusses the network connectivity of USF's data centers, which host critical business data and applications, and the redundancy measures in place, such as primary, secondary, and disaster recovery data centers. It emphasizes the high-speed connections within the data center and the network's design for reliability, including dual power supplies and connections to different switches for critical servers.
🛠️ Management and Monitoring of USF's Enterprise Network
The second paragraph delves into the management and operation of the nearly 10,000 devices that make up USF's network by a small team of experts. It outlines the roles of network engineers and operation staff in designing, supporting, configuring, and installing network equipment, as well as maintaining physical cabling and telephony systems. The paragraph also describes the extensive monitoring practices in place, which include checking device health, traffic levels, and performance statistics every three minutes. This monitoring aids in capacity planning, troubleshooting, and security. The discussion concludes with the distinction between enterprise-class and SOHO-class equipment, emphasizing the need for high-end features, reliability, and remote management capabilities in large networks like USF's.
Mindmap
Keywords
💡Desktop
💡Networking
💡Wang Routers
💡Backbone Switches
💡Internet Service Providers (ISPs)
💡Border Gateway Protocol (BGP)
💡Data Centers
💡Ethernet Connections
💡Enterprise Class Gear
💡Network Monitoring
💡Capacity Planning
Highlights
The importance of a desktop being able to communicate with other devices on the network for effective networking.
Network diagram showing the connection of Wang routers to backbone switches for wide area connectivity.
USF's Internet connectivity is provided through multiple service providers including Cogent, Time Warner Telecom, and Florida Lambda Rail.
Ethernet connections over single-mode fiber are used for delivering Internet links to USF.
LAN routers use BGP to determine the best ISP for routing traffic to its destination IP.
Approximately 450,000 routes are advertised by USF's ISPs.
The Tampa data center's direct connection to the campus core and its role in hosting USF's business data and applications.
USF maintains multiple copies of its data across different data centers for continuous service.
Servers within the data center are connected at 1Gbps and 10Gbps speeds.
Redundancy in network design to handle switch failures and maintain server connectivity.
Network engineering group's use of software to monitor device health and availability every three minutes.
Monitoring of various network statistics for capacity planning, troubleshooting, and security.
The requirement for enterprise-class gear in large networks like USF's for reliability, security, and manageability.
Differences between enterprise-class and SOHO-class gear in terms of features, reliability, and scalability.
Enterprise-class equipment's ability to handle heavy traffic loads and remote management capabilities.
The small group of experts managing and operating the entire USF network of nearly 10,000 devices.
The network infrastructure's design from the desktop to the campus core and out to the Internet.
Discussion on the types of networks required for an enterprise-class network.
Encouragement for viewers to continue learning about network technologies and consider working in the field.
Transcripts
so far we focused on how a desktop is
connected to the usf network but just
having a desktop on the network isn't
useful unless it's able to communicate
with other devices like servers and
other clients after all the point of
networking is to enable communications
so we'll now look at how traffic gets to
the Internet once it reaches USS campus
core looking again at the high level
network diagram you can see that there
are a pair of Wang routers connected to
the backbone switches these devices when
one in way and to provide wide area
connectivity to USF's Internet service
providers when one is a Cisco ASR 9006
and when two is a Cisco 7600 usf obtains
internet connectivity through cogent
time warner telecom and florida lambda
rail these links are delivered to USF as
Ethernet connections over single-mode
fiber this is the same kind of Ethernet
we've used everywhere along the path
from our example desktop traffic is
routed to the land routers by the campus
backbone and those routers determine the
best ISP to use for reaching the
traffic's destination IP in order to
make this decision the LAN routers run
BGP or border gateway protocol there are
currently approximately 450,000 routes
advertised USF from our ISPs once
traffic is forwarded to the appropriate
ISP routers in their network and many
other networks beyond that ISP deliver
the traffic to its destination when that
remote server responds and sends traffic
back to the desktop at USF the same BGP
routing tables guide the traffic through
the internet back to USF and our network
delivers it back to that machine we will
now look at the network connectivity to
the servers USF runs in its data centers
these data centers hosts all of us s
business data and applications such as
student assignments grades email
websites employee payroll billing travel
etc to ensure continuous service USF
maintains multiple copies of its data
USF houses a primary data center on the
Tampa campus a secondary data center in
Winterhaven in a disaster recovery data
center at Clemson University in South
Carolina the Tampa data center is
directly connected to the campus core as
shown in the high level
or diagram to Sisko 65 hundredths
provide connectivity to all the servers
and devices in that datacenter
individual servers within the data
center are connected at 1 gigabits per
second and 10 gigabits per second speeds
these servers connect to access layer
switches which connect to the 6500
really the network gear in the data
center isn't that much different than
the equipment BC connecting desktops in
buildings for example all switches
providing connectivity to the servers
have dual power supplies connected to
different electrical circuits all
critical servers have network
connections to two different switches so
that if a switch failure does happen the
server can send traffic through its
other switch each switches uplink to
different Cisco 6500 so that if one of
those 6500 s fail connectivity would
continue through the other with nearly
10,000 devices making up USS Network
failures are inevitable so the network
engineering group runs software to
monitor every one of these devices we
check their availability and health
every three minutes we check things like
CPU memory utilization optical power
levels on the fiber links room
temperatures ups capacity and general
functionality of the device if any of
these parameters are out of normal
values we are notified and someone from
the operations or engineering group will
troubleshoot the issue we also monitor
various statistics from our network here
we graph input and output traffic levels
for every switch port and router port
concurrent users and throughput on every
wireless access point network
performance statistics across our
backbone and inter-campus LAN links and
call volumes on our telephony systems
basically we monitor as much information
as we possibly can and attend to any
unusual observations these statistics
help us with capacity planning
troubleshooting and even security for
networks the size of USF's to function
the equipment used and then must be
reliable secure feature-rich remotely
manageable and have an assurance with
technical support for the manufacturer
it must be what's referred to in the
industry as enterprise class gear
equipment from vendors like Linksys the
link and Belkin is typically Soho class
small office or home office class gear
these low-end devices work well in the
home or in a small office that's not
part of a larger enterprise network they
have many useful features that work in
environment for example they are
inexpensive and usually require no
configuration and are ready to run out
of the box when they fail they just need
to be replaced with similar equipment
for an office with five to ten users
this is adequate but they do not do well
in a large network like we have here at
USF USF needs many high-end features not
available in solo class gear and we need
the ability to remotely monitor and
manage all aspects of each networked
device enterprise class equipment is
certainly more expensive than Soho gear
but that extra cost helps guarantee the
networks we build provide the level of
service our users demand they are
capable of handling heavy traffic loads
from thousands of users their remote
administration and configuration
features allow a small group of network
experts to manage thousands of devices
their configurability allows USF to
respond to changes in network feeds by
reconfiguring traffic routes as
appropriate the entire USF network of
nearly 10,000 devices is managed and
operated by a small group of experts
five engineers design and support all of
the networks and services for network
operation staff handle tasks such as
equipment configuration and installation
we also have several staff maintaining
the physical cabling handling requests
from users for ads moves and changes and
we have a team who handles telephony for
the campus us scene USS Network from the
desktop all the way up to the campus
core and out to the internet you've seen
how traffic gets to the internet and how
the infrastructure in u.s. s primary
data center is designed we've discussed
how all of this infrastructure is
monitored and what kinds of network
you're required for an enterprise class
network I could literally spend hours
talking about any one of the topics
we've covered in this video but
hopefully this short introduction has
provided you a good understanding of the
technologies which make up the USS
Network the network of any large
enterprise will share many of the
similarities with this network and if
you find these technologies as
interesting as I do
I hope this has encouraged you to
continue learning about them and perhaps
even consider working on them
Посмотреть больше похожих видео
5.0 / 5 (0 votes)