The networks that carry Internet, other network, and telephone traffic worldwide are truly global in scale. Comprised mostly of fiber-optic cables laid under ground by 'long line companies' interconnect Internet and Telephone Exchanges all around the world. In a fiercely competitive communications marketplace, there is more 'bandwidth' available for less cost than ever before and it keeps getting faster and cheaper.
Check out this detailed gallery about Submarine Cables that carry Internet traffic around the world. Here's a Spanish (or Portugese?) language page from Geografia e Geoolitica that shows land-based fiber-optic networks, too. Although fiber-optic cables are mostly immune to the EMI (Sunspots, solar storms, thunderstorms) that plagues copper cables, they do require periodic maintenance and repair.
I used to joke about sharks biting undersea cables, now after reading this Slate article I speak seriously!
Here's a Built Visible Article about undersea cables.
'The Internet Backbone' is provided by big telecommunications providers referred to as Tier 1 Networks. These companies 'peer' with each other inside IX-Internet Exchanges by interconnecting their industrial-strength routers with fiber-optic jumpers inside these huge facilities. Since they own the WAN media, Tier 1 providers don't have to pay anybody else for network bandwidth, but they are required to connect with competitors' networks. Tier 2 and 3 ISPs may co-locate with Tier 1 providers and also have facilities in their areas of service. They pay Tier 1 providers for their circuits and bandwidth, and their customers, residential or commercial, pay them for internet service.
This is an introduction to Internet, Ethernet, and 'last-mile' services that connect SMB-Small to Medium Business and larger organizations to the public switched telephone network and their internet service providers.
Networking in the 20-teens is easier than ever! Everything runs thru a few kinds of boxes and protocols, and networking is highly standardized all over the world. The Internet, The World Wide Web, private internets, and LANs-Local Area Networks use IP Routers and Ethernet Switches and a handful of protocols we call 'internet' and 'ethernet'.
Networking from the 1970's was a very expensive Tower of Babel with dozens of VANs-Value Added Networks that charged as much as $1 to deliver an EDI-Electronic Data Interchange document like a Purchase Order, Confirmation, Advanced Shipping Notice, Manifest, or Invoice from one trading partner to another. On-line services like AOL, Prodigy, Earthlink and lesser-known 'bulletin boards' depended on dial-up connections, often long-distance, and e-mail was not transferred among services. This was _not_ The Internet.
Since the '90s, The Internet has mostly replaced the VANs and these documents are now transmitted 'for free', securely, among trading partners to facilitate today's JIT-Just In Time eBusiness. ISPs replaced many early on-line services as The Internet emerged. Successful on-line services migrated their customers to The Internet, facilitated their browsing and email on the World Wide Web, and kept many of their customers through today.
Along with these several networking and internetworking protocols come a few encryption schemes to make internet connections into VPNs-Virtually Private Networks. SSL-Secure Socket Layer, TLS-Transport Layer Security, PKI-Public Key Infrastructure, SSH-Secure Shell, and other security protocols provide 'virtual privacy' in an open environment where every packet must be accurately addressed or it will not be delivered. Encrypting internet, web, and email is secure enough for most purposes. Where the NSA can decrypt 50K+ SSL transmissions a day, it's prohibitively expensive for ordinary crooks and crackers...
IP Routers, Ethernet Switches, Ethernet Hubs -- OSI Layers 3, 2 & 1 The 7-Layer OSI Model- Open Systems Interconnect is the most-used model to describe the hardware and software for networking. There is also a 5-Layer Internet or TCP model. These are not hard, engineering standards. They are descriptive models used to discuss networking in the decades after proprietary networks faded from use.
Routers are OSI Layer 3 devices that use IP addresses (IPV4 or IPV6) to connect at least two networks together. Switches connect devices in a LAN, which may include wireless.
A LAN may connected to other networks, and/or to The Internet by connecting an IP Router to the LAN's switch.
Ethernet Switches are OSI Layer 2 devices that use MAC addresses to deliver Ethernet 'frames', which may contain Internet packets, among routers, computers, and printers attached to the LAN.
Ethernet Hubs operate at OSI Layer 1 and are almost a historical note in 2016. They were at the center of the ordinary LAN through the '90s and into the '00s, then were eclipsed by Ethernet Switches during the '00s, where switches' prices rapidly decreased from about $150 per port in '99 down to a few $ per port in the mid '00s.
Ethernet Hubs were a huge improvement over the coaxial cables and DB-style connectors used in early networks. using RJ45 connectors and Cat5-6 cabling to provide easy and reliable 'physical media' relative to old BNC, Coax, TwinAx, and DB connectors. After Ethernet and Cat5 were widely adopted, the costs of networking equipment and premesis wiring plummeted in a new, highly competitive marketplace where proprietary standards were eschewed.
Ethernet Hubs are simple 'repeaters' that echo all signals put to the bus to every connected device so that every device 'sees' or 'hears' all traffic on the LAN. This can contribute to security vulnerabilities since it is fairly simple for somebody on the LAN to run 'sniffer' software and capture ethernet packets destined for somebody else.
Every device on an Ethernet Hub or Coaxial cable can hear every other device, so CSMA/CD works for traffic management. This is 'Carrier Sense Media Access with Collision Detection'. On a wired ethernet a node with a packet to transmit listens to the carrier frequency on the media and waits for quiet before transmitting the packet. As the packet is transmitted the device listens for the 'echo' on the frequency. If the echo exactly matches the packet transmitted the node waits for confirmation of the packet; if a 'collision' with another device's transmission is echoed, all the nodes with packets to send calculate a 'random backoff interval' and the one with the shortest interval transmits and listens...
If an Ethernet is not overloaded this CD-Collision Detection scheme works well enough. An Ethernet will fail if there are too many devices with too much traffic on the network. At times of peak usage the QoS-Quality of Service for each user may be very poor. QoS on an Ethernet is determined by the number of users and servers and their demands for bandwidth at the moment. CSMA protocols are 'stochastic', with traffic managed by listening for collisions and random backoff intervals. Modern Ethernet is very quick and 'rule of thumb' measures are easy for network managers to apply to ensure adequate QoS and satisfied network users.
Where QoS is critical, Token Ring or other 'deterministic' traffic management technology such as polling or multiplexing should be considered.
Modern Ethernet Switches
WiFi-Wireless Ethernet WAPs-Wireless Access Points can be thought of as 'wireless switches' that operate at OSI Layers 1 _and_ 2. WAPs and WiFi clients are Layer 1 devices where the radio frequency behaves much like a bus: each client device on a WAP's frequency can 'hear' other devices within its range, but there may be devices at the opposite edges of the WAP's range that cannot hear each other's transmissions so collisions at the WAP would be frequent. WAPs operate at OSI Layer 2, employing MAC addresses in CSMA/CA's RTS-Request to Send/CTS-Clear scheme to avoid collisions on the range of frequencies attended by the WAP. CSMA/CA is 'Carrier Sense Media Access with Collision Avoidance'.
A typical router for use in a home or SMB-Small to Medium Sized Business may be combined with a WAP-Wireless Access Point and/or Ethernet Switch. This makes it very easy to connect, where there is one connector for the WAN media and the WiFi is built-in. Most of these hybrids also have one or more connectors to a wired ethernet.
These small IP routers, or modems, seldom connect to more than one internet medium and may function as IP bridges with only one choice of route. Lots of SOHO and residential networks don't have a router at their premises. 'The box' that provides internet connectivity may be a cable or DSL-Digital Subscriber Line modem attached via a serial link to an ISP's router.
Where a small router in a SOHO or SMB only connects to one WAN and one LAN, an 'industrial strength router' may connect to dozens of WANs and dozens of LANs within a NOC-Network Operations Center or IX-Internet eXchange.
An ISP or larger organization may use one or more large routers to interconnect its networks with The Internet via more than one ISP. For example, VCU uses a dozen industrial strength IP routers, each connected to all others on campus via fast optical circuits maintained by VCUNet (look for VCUNet on some manhole covers as you walk around campus). There are a few, redundant, large IP routers that attach to several OC-Optical Carrier circuits to Tier One ISPs. Most universities are blessed with huge, enterprise class bandwidth.
Google 'Brocade Router' or 'Juniper Router' to get a look at some competition for Cisco Routers. Cisco has been the leader in internetworking technology and continues to be recognized as a giant. But, in this very open tech market Cisco has lots of price competition from these other manufacturers. A large router may cost $20K or lots more depending on its capacity.
Sketch on the board and discuss traffic management and practical applications of the basic Network Topologies: serial/point-to-point, bus, ring, star, tree/hierarchical, mesh.
Terms introduced: Polling, RTS/CTS, Terminator resistors, Trunk, branches, RS232, SDN-Software Defined networks, Points of failure, VLAN-Virtual LAN, Etherfabric, CSMA/CD, CSMA/CA, Clustering, Heartbeats...
Chapter 4 in the text, and the early chapters in any network certification guide, discuss network topologies and methods for traffic management on networks. Network topologies are covered in the text and in many on-line references. Wiki: Network Topology is a good intro.
Traffic Management on Ethernets: Ethernets use bus topology where all devices have a unique MAC address and are connected to the same bus, which may be copper wire, optical fiber, or a radio frequency. Wired Ethernets use 'collision detection', CSMA/CD, to manage traffic -- all the devices connected to the copper Coaxial cable or Ethernet Hub can 'hear' each others' transmissions so CSMA/CD works. It works well enough that Ethernet is the most used type of network. When a device is ready to transmit a packet it 'listens' until the carrier is clear, transmits the packet, and listens for a 'collision'. If there is no collision the sending device waits for an acknowledgement of the packet. If there was a collision, all devices with a packet to transmit calculate a random 'backoff interval' and the device with the shortest interval attempts again to transmit its packet. In a lightly loaded Ethernet collisions are rare and throughput is high, although it is never anywhere near the bandwidth available In an overloaded Ethernet QoS quickly suffers and throughput is only a small fraction of the bandwidth.
Modern Ethernets usually use Ethernet Switches rather than Hubs. These devices have a bus at their heart that runs 10 or 100 times as fast as the ordinary ports, and the 'collision domains' are better managed. The bus in an Ethernet switch is connected to the ports through a 'switching matrix' of circuits and buffers that greatly improve performance and enhance security relative to the old Ethernet hubs.
WiFi uses Wireless Access Points to connect wireless devices to an Ethernet. Most residential internet service by DSL, Cable, or fiber includes a 'wireless router' and many residences don't use wire at all. Wireless Ethernet uses 'collision avoidance' to manage traffic: CSMA/CA. Wireless devices at opposite, extreme edges of the WAP's range may not be able to hear each others' transmissions so collision detection will not work. The WAP actively manages traffic by listening for RTS-Ready To Send signals from the wireless devices with packets to transmit and sends a CTS-Clear To Send signal to the device it chooses for the next transmission. The RTS/CTS of CSMA/CA uses more of the LAN's bandwidth to manage traffic than CSMA/CD.
Both CSMA/CD & CA involve 'listening' and an element of chance for the random backoff interval following a collision. They are stochastic. QOS-Quality of Service plummets at peak periods of use if an Ethernet is not sized properly. Maximum sustainable throughput for devices on ethernets is always much less than the bandwidth. But, Ethernet works well enough that it is the most common type of network worldwide. Where nothing should be left to chance, as for surgical networks or process control, Ethernet is not the best choice for traffic management.
Traffic Management on Rings is often by 'Token Passing': Token Ring networks are a variety of ring topology that uses Token Passing for traffic management. Token Rings are 'deterministic', not stochastic. A node on a token ring transmits only when it gets the token, so QoS can be engineered to handle peak traffic and there is no contention for bandwidth. The 'token' is an empty packet the nodes pass among themselves. When the token arrives, the node transmits packets for it's alloted time, then passes the token to the next node.
ARCNET is a bus network that uses token passing. It was a predecessor of Ethernet and remains in use today. Where Ethernet can be flooded and gives poor QOS at peak demand, ARCNET QOS remains stable, allowing a slower ARCNET (2.5 mbps) to out-perform a faster Ethernet (10 mbps). Ethernet prevailed in the market as it got to 100 mbps and was more flexible than an ARCNET, allowing a tree/hierarchical arrangement of hubs or switches where ARCNET is limited to a bus/coaxial cable topology.
Serial Networks use star topology where each remote device has a dedicated, serial, connection to a central unit. Some versions of serial networks 'poll' their devices and avoid contention all together. Or, they may respond to 'interrupts' or RTS signals from devices so bandwidth is not wasted polling idle devices.
Printers, lab equipment, scales, and other serial devices that would be cabled directly to the serial ports on a mid-range computer in the past may now connect to 'serial gangs' in expansion slots on server-class machines, or to 'termservers' that connect the serial devices via Ethernet or ArcNET. There is usually no address for the peripheral devices in Serial Networks -- each device takes the address number of the port they are jumpered into.
A DSL or other leased circuit that connects to an ISP or branch office may be referred to as 'the serial link' since there is no choice of route. PLC-Programmable Logic Controllers for operating manufacturing equipment and utilities like dams, barrages, or power stations often use RS-232 serial networks to connect to sensors and controls, as do audio-visual systems for classrooms and business.
Far from obsolete in spite of decades and decades of service, The Internet's suite of protocols includes SLIP-Serial Line Interface Protocol and PPP-Point to Point Protocol for more efficient operation of serial circuits that are switched or fixed, vs. packet-switched circuits of Ethernet and Internet.
Traffic Management on Internets is more complex since there may be multiple routes involved, where only one route is available on the prior network topologies. Routers connect two or more networks. A common arrangement is for a SOHO LAN to connect via a 'DSL Modem' or 'DSL Router' in the home or office via a 'serial link' (copper wire) to a port on a DSLAM-DSL Area Multiplexer located in the neighborhood or in the ILEC's facility. The DSLAM connects to 'industrial strength' routers in the DSL provider's IX-Internet eXchange, which connect to multiple fibre circuits on the Internet Backbone, which connect to other routes across The Internet. We want our ISP to be well-connected to lots of high-speed fibre, and we don't want the to oversell
Routers are 'gateway devices' that dispatch packets of data through an internet via the best available route at the instant each packet arrives for transmission. Routers adapt to 'line conditions', busy circuits, circuits that go dead, or circuits recently provisioned, to ensure quickest, error-free deliver of packets.
On the 'inside', LAN side, of the gateway, Routers use ARP-Address Resolution Protocol to match up IP addresses with MAC addresses on the LANs they service. On the 'outside', internet side, routers use RIP-Router Interface Protocol to discover neighbouring routers in The Internet, or an internet, and other protocols (sometimes proprietary) to gather metrics about routes over the horizon. Routers' operating systems dynamically supply Dijkstra's Algorithm with the best data to make the choice of route via OSPF-Open Shortest Path First and other methods. They use IP-Internet Protocol to move packets from router to router on routes provisioned by TCP and other handshaking protocols. An IP router's 'routing tables' are a mix of hand-edited routes, sometimes to reflect commercial agreements, and dynamic routes kept fresh by the router's OS.
TCP handshakes to establish a 'connection' between the end points. IP manages the traffic of packets on the connection, involving error-detection & correction and sliding-window flow control to ensure accurate transmission of data.
Ethernet Fabric is LAN technology that provides multiple routes for network traffic. These are used in SAN-Storage Area Networks and data centers. New ARM-64 SoC components include Ethernet Fabric and are optimized to participate in Spine and Leaf networks.
These next links are about most types of network media, the buildings they connect, and the services that run over them. There is some redundancy in the next two links and anything that's repeated is right important...
This links to discussion of Networking Infrastructure: Ethernet Hubs & Switches, IP Routers, Internet Backbone. Alphabet Soup: PAN, Bluetooth DID, FCC, CE & IC, LAN, MAC Address, MAN, WAN, IP Address, Private IP Address, Domain Names & DNS, ISP, AS, RIR, ARIN, IEEE, ISOC, IETF, W3C, ICANN, SSL, CA, LEC, CLEC, ILEC, NAT, RIP, ARP, PSTN, POTS, ISDN, T1, T3, OC3 thru OC192, DSL, &c...
This is a classroom demo, showing structured premises wiring components, equipment likely to be in a network room, and some test equipment. It didn't happen Spring 16, google is your friend to fill in the details...
'Storage' is one of the essential components of information technology, right up there with Computers and Networks. For decades magnetic HDD-Hard Disk Drives have been the ordinary storage technology, and SSD-Solid State Drives are quickly finding their way into the legacy as they've become about the same cost as HDD, with several benefits including speed of access and reliability of solid-state parts vs. rotating disk mechanism.
Attachment of Storage: Storage is attached to computer systems in several ways: DAS-Direct Attached Storage, NAS-Network Attached Storage, SAN-Storage Area Networks, and modern web-scale and hyperconverged storage systems that manage redundency on a global scale. Atttachment is discussed later.
Disk Geometry & Physics introduces an alphabet soup of abbreviations for concepts required for management of storage on disks: IDE, CHS, LBA, Clusters, Slack Space & Fragmentation, ZDR, &c...
Most disks in a server environment are deployed in RAIDs - Redundant Array of Independent Disks, not one at a time. RAIDs increase the reliability of HDD and SDD and can also boost read/write performance relative to stand-alone disks.
Enterprise doesn't run on cheap disks! The ordinary, $79/Terabyte HDD would be beaten to death in a busy database server where the disks are exercised hard 24X7X365. SAS - Serially Attached SCSI drives cost a few times more, spin nearly 2X faster, and are engineered for 15 years MTBF and heavy use. SSD are about the same price per TeraByte as HDD in 2016 and will eclipse HDD technology and make it obsolete in the foreseaable future...
Appropriate transaction logging and backup schemes can ensure as close to 100% durability of data as is possible, beyond 99.9999%, so that a server can be restored to the point of failure following a data disaster. Modern techniques can provide quick or 'seamless' recovery in the wake of equipment failure or network room or regional disaster so that system users may not be aware that there was a failure. Without backup and transaction logging, systems may not be recovered after a disaster. The cost of system failure without adequate backup can be so great that the business fails in most cases where there was not adequate backup.
Read on for more stuff about management of storage technologies...
In 2015 it was revealed that the NSA, and our enemies, can embed Spyware in the firmware of hard disks. This technique has long been involved in the plotlines of spy novels -- nearly as long as there's been IDE - Integrated Drive Electronics there have been ways to exploit the 'little on-board computer' on each drive in the supply chain, or by malware that flashes itself into the IDE's firmware. Today, this practice may be more widespread and vigilance is required, including audit of firmware in regular system health checks.
Why do we backup data?
The easy answer is 'to continue or recover business after a system disaster'. More than half of businesses that lose their computer system without a good backup fail.
The real answer must include 'and, to continuously prove the integrity of data'. If an organization is lucky there will never be a system disaster. But, backup sets and transaction logs will be used every day to audit and prove the integrity of data and investigate irregularities.
Backup sets and transaction logs support the 'I' in the classic Information Security triad CIA: Confidentiality, Integrity, and Availability. Without regularly examining backup sets and transaction logs and comparing them to the on-line records it can be impossible to detect or prevent loss or theft of data and impossible to get it back.
No organization wants a customer, employee, supplier, or the tax man to show them records produced by their system that they can't explain. That would demonstrate a lack of integrity and cast doubt and suspicion on all past and future dealings.
Hardware failure and local or regional disasters are _not_ the reason for most data disasters requiring recovery from backup media and transaction logs. Human error, ineptitude, or malice are much more likely.
Maybe somebody working on a system puts a semi-colon where a comma should be in an SQL update statement and accidentally wipes out the table holding all the transaction data for the past few hours. Or, a consultant demonstrates 'SQL Injection' thinking he's working on a development system and wipes out the production database. Or, a cracker finds his way into a system, has his way with it, and wipes it clean when he's done. Or, there is nobody watching files that grow and one grows so large it eats the file system...
Hardware failure and disasters in a network room, building, locality or region must be considered. Even if they're not as likely as human malice or failure they do occur. Here are some good reasons Why We Backup Stuff, negative examples of how to mitigate the risk of data center disasters. Here's another good look at Database Disasters. Tom's IT Pro is an excellent resource for real-world tech, including backup.
Components of backup:
In the event a recovery is needed: the hardware is prepared and the operating system is restored; data from the last full backup is restored; data from incremental backups is restored; recovering data on transaction logs brings the system back to the point of failure.
Modern 'de-duplication' techniques as engineered by companies like IBM or Barracuda can provide reliable copies of every version of every record without duplicating all the un-changed records, too. Outsourcing tape backup to a company that uses tape-storage jukeboxes or robots with or without de-duping is a good option as companies turn to IaaS-Infrastructure as a Service.
The sun never sets on some enterprises, so the system may never be quiet for a backup. One use of virtual servers is so that a multi-national organization can run a system with the clock set for each time zone. Companies like IBM, Oracle, or Barracuda can engineer a solution so that backups and can be taken while the system is not quiet and can be used with transaction logs to recover a system to any point in time.
RAM is the key issue in these 64-bit days where a server-class machine can reference as much RAM as midrange and mainframes could at the turn of the millennium. We already have server-class machines that can handle a TeraByte of RAM and 24 or 48 Cores. Access to data in RAM is upwards of a couple hundred thousand times faster than access to Disk! RAM speed is expressed in NanoSeconds, DISK's relative sloth in MilliSeconds.
A problem with having huge RAM 'all in one basket' like a server-class machine is that if the machine fails all the data in memory is lost, affecting dozens or hundreds of customers' or employees' orders or work.
Big machines, midrange and mainframe (but not servers yet), can hold _ really huge_ RAMs of a few or several TeraBytes in their big chassis, so they gain a huge speed advantage by keeping users' active data and programs cached in RAM. To mitigate the risk of a memory unit failing, the midrange and mainframe machines have mirrored or 'RAIDed' RAMs that allow them to rip along with one RAM if the other develops errors. RAM modules can be replaced without taking the system down.
There are several types of RAM from the fastest, volatile, static ram on a mainboard to much slower, non-volatile memory on a USB-Drive, flash memory array, or SDRAM card. Here's a good Fast Guide to RAM that discusses the differences.
Ch. 2 in the Text covers the hi-level topics of computing hardware, the above link is a survey of legacy and emerging eBusiness tech.
Computer hardware does nothing without software. Some of the software may be packaged as 'firmware', which is software burned or flashed into a ROM, EPROM, EEPROM attached to a mainboard or built into a SoC. Software doesn't do anything without hardware, either. Remember, a 'platform' is the combination of CPU/hardware and OS/software.
Virtualization gathered momentum as the new millennium rolled on and is now an ordinary feature of many computer systems. It is applied in a few different ways from 'virtual servers' where one big machine can behave like several servers, each with its own OS and authentication scheme.
For decades a fundamental security step has been to run 'a server dedicated to each service'. This helps secure a system by making it more difficult for a cracker to exploit a vulnerability in one service to gain access to all of them. A website might have different servers dedicated to firewall, database, applications, and web. Today, that is more likely to be implemented as 'a virtual server instance' for each service. Where firewalls in the past were often discrete appliances dedicated to fire-walling, today firewall's are likely to be virtual.
'Virtual desktops' have become more common. Instead of each employee having a powerful desktop or notebook computer where they must sit to access their applications and data, their 'desktop' runs as a virtual instance on servers dedicated to the purpose. Then, employees can access their desktop from anywhere with some bandwidth and a browser. The employer benefits in several ways, not the least of which is that network techs seldom need to visit an employee's desk to install or upgrade software.
Citrix XenServer, Microsoft Hyper-V, Oracle VM, VMWare VSphere and other softwares including open source alternatives compete for this growing market. Today's larger workstation/server machines with large RAM and many cores help make virtual desktops a good alternative.
Many network managers believe that 'virtual firewalls' are better than the real ones. Here's an article at TechPageOne.com about Managing Risk with Virtualization -- it addresses planning for disruption, minimizing downtime, maximizing continuity, and facilitating recovery. 'Virtual Networks', VLANs, behave like switches and routers but are implemented within a big machine's bus.
Hyperconverged systems and Web-scale IT are modern patterns for management of real and virtualized system resources.
2016 Fall: Beyond the cutoff point for Quiz #2 is not accurate to be carried into Quiz #3. If you're looking for Quiz #3 questions please check: Quiz #3 Sample Questions