A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
Active Memory: Task specific technology for packet processing from MUSIC Semiconductors, it is the next step beyond conventional CAM (Content Addressable Memory). Active Memory interleaves memory with proprietary MUSIC logic to deliver hardwired funtionality and performance unmatched by devices using SRAM, DRAM, standard CAM or software alternatives.
address: The number of a particular memory or peripheral storage location. Like post office boxes, each byte of memory and each disk sector has its own unique address. Programs are compiled into machine language, which references actual addresses in the computer.
ADSL: Asymmetric Digital Subscriber Line: Modems attached to twisted pair copper wiring that transmit from 1.5 Mbps to 9 Mbps downstream (to the subscriber) and from 16 kbps upstream, depending on line distance.
algorithm: A step by step mathematical method of accomplishing a task. A sorting algorithm may be like this: 1) Progress down your list from the top; 2)When you encounter something that is less than the last item swap the two; 3) Repeat 1-2 until you reach the end of your list.
application: A program. Programs such as Microsoft Word, Excel, Photoshop, etc. are applications. Applications give your computer instructions that provide you with tools to accomplish a task. For instance, a word processing application provides space to type in, means to format the text, a spelling checker, etc.
Application Programming Interface (API): APIs allow you to program to a pre-constructed interface (the API) instead of programming a device or piece of software directly. This allows for faster development, since programming to a device's API is designed to be easier than programming to a device directly. APIs allow you to program without having intimate knowledge of the device or software that you are sending commands to.
Application Service Provider (ASP): An ASP is a company that provides remote access to applications, typically over the Internet. ASPs are used when an organization finds it more cost effective to have someone else host their applications than to do it themselves. The applications served up can be as simple as access to a remote fileserver, or as complex as running a word processor through your browser.
Application-Specific Integrated Circuit (ASIC): An ASIC is a circuit designed for a very specific purpose, such as the processors in PDAs or the chips on a motherboard chipset. ASICs contrast with more general-purpose devices such as memory chips or x86 processors that can be used in many different applications.
ASSP (Application Specific Standard Product): A standard product that has been designed to implement a specific application function, as opposed to a general-purpose product such as a RAM.
Asymmetric Digital Subscriber Line (ADSL): This is a new technology that the telecommunications company's answer to cable modems. It supports data speeds over 2 Mbps downstream (to the user) and slower speeds upstream (to the Internet). This technology should eventually supplant ISDN if it achieves a reasonable price point.
ATM: Asynchronous Transfer Mode (ATM) - A networking protocol designed to move multimedia data around with high reliability and speed. Some ISPs use ATM as the protocol for their backbones. Asynchronous transfer uses start and stop bits for every bit of data transmitted, making it secure enough for financial transactions. (ATM does not stand for Automatic Teller Machine) Only until higher and faster transmission speeds were capable of handling all the additional security bits did asynchronous transfer become possible. (See also synchronous)
back to top
backbone: A segment of a network that's often a higher speed than the rest of the network and connects all the other segments. If you don't have a fast backbone, your network will lag. Many ISPs are constantly restructuring their backbones for this reason.
backplane: As opposed to standard cabling schemes where flexible wires are used, a backplane refers to a rigid circuit board that will support higher connection speeds and more logic. An example of a backplane is in network connection devices, such as large enterprise scale switches and routers. Some of them have a high-speed backplane, and you can plug a group of slower network connection devices into the high-speed backplane.
bandwidth: The amount of data transmitted per unit of time, typically bits per second (bps) or bytes per second (Bps).
How much stuff you can send through a connection. Usually measured in bits-per-second. A full page of English text is about 16,000 bits. A fast modem can move about 15,000 bits in one second. Full-motion full-screen video would require roughly 10,000,000 bits-per-second, depending on compression.
BGA (Ball Grid Array): As opposed to a pin grid array, a ball grid array is a type of microchip connection methodology. Ball grid array chips typically use a group of solder dots or, balls, arranged in concentric rectangles to connect to a circuit board. BGA chips are often used in mobile applications where PGA chips would take up too much space due to the length of the pins used to connect the chips to the circuit board.
binary search: This is a search technique that splits an ordered list into halves until a result is found. Thus it splits the list into a top and bottom half, then picks the half that the search term is located in (which can be determined because this search is used on ordered lists) and splits that in half, and so on until the search term is reached.
bps (Bits per second): This is generally a measure of how fast some device communicates, usually in thousands of bits per second (Kbps) or millions of bits per second (Mbps).
bridge: Bridges link two or more network segments that use the same network technology. The network topologies do not have to be the same, i.e. you can bridge Ethernet and Token Ring. If the protocols are not the same on segments, then you've got a router. Bridges are used to reduce network traffic, increase cable distance (by acting as a repeater), and also to link dissimilar network topologies that run the same network protocol.
broadband: This is a technology that refers to the transfer of multiple signals over a single medium. In slang terms, it is any Internet connection that allows for higher transfer speeds than an analog modem, most often applied to cable modem access. However, it is sometimes used to refer to DSL, satellite and wireless Internet services as well.
broadcast: This refers to a method of sending information over a network. With broadcasting, data comes from one source and goes to all other connected sources. This has the side effect of congesting a medium or large network quickly. Sometimes, broadcasting is necessary to locate network resources, but once found, more advanced networking protocols change to point to point connections to transmit data. Nowadays, switches and routers often do not pass along broadcast packets, but in the days of shared Ethernet, broadcasting could really congest a network with packets quickly.
buffer: This is a temporary location to store or group information in hardware or software. Buffers are used whenever data is received in sizes that may be different than the ideal size for the hardware or software that uses the buffer. Another use of buffers is to keep hardware from getting overwhelmed with information. In that scenario, you use a large buffer to hold data until a device or program is ready to receive it.Buffers must be optimized in size to work efficiently for the purpose they are designed.
byte: Simply put, it's 8 binary (0 or 1) bits. A string of 1s and 0s that represents a number from 0 to 255. 256 combinations represent 2 to the 8th power. For example '01100101' is one byte of information, and represents one of a possible 256 values.
back to top
C Language: This is a programming language developed in the late '70s. It became hugely popular due to the development of UNIX, which was written almost entirely in C. C was written by programmers for programmers and it lets you write code in ways that more structured languages do not.
cache memory: Generally a small chunk of fast memory that sits between either 1) a smaller, faster chunk of memory and a bigger, slower chunk of memory, or 2) a processor and a bigger, slower chunk of memory. This is to provide a bridge from something that's comparatively very fast to something that's comparatively slow. Most computers have cache memory that holds some of the information from main memory. When the processor needs the information it takes it from the speedy cache instead of the slower main memory. Cache memory GREATLY increases the speed of a computer by storing data that is most often accessed.
CAM (Content Addressable Memory): A CAM device is a semiconductior in which packet address information is stored such that a destination address may be located by making a direct match. By comparing what is sought for against all pre-stored possibilities, searching through a table is eliminated and routing of packets is expedited. Having this functionality hardwired in has proven to be a faster method than software routines, search engines and hash tables performing a similar function. CAM has emerged as the search engine of choice because of their ability to conduct high-speed table lookups that are required to meet the demands of ever-growing traffic on the Internet.
CAM processor: Helps increase the speed of packets through switches and routers
CAM-based IC's: Content Addressable Memory - Integrated circuit. CAM based electronic circuit in which many active or passive elements are fabricated and connected together on a continuous substrate, as opposed to discrete devices, such as transistors, resistors, capacitors and diodes.
capacitor: This is an electronic component that stores up an electrical charge to a certain level, and then releases it. A capacitor stores energy between two conducting plates separated by an insulator, or dialectric. Capacitance, or the amount of current that is stored is measured in farads. You may notice large looking electronic cylinders on your motherboard. These are capacitors, and they are used to provide clean power to your processor. You'll also find larger capacitors in your power supply. Television and CRT monitors contain many capacitors, that's why you should never open up a TV or CRT even when it is off, as there may still be a dangerous electric charge in the capacitors.
cascadable: The ability of hardware to phyically link togther to extend storage capacity.
circuit switching: A switched circuit is only maintained while the sender and recipient are communicating, as opposed to a dedicated circuit which is held open regardless of whether data is being sent or not.
client: This is the general term that can mean a few things. In the client/server architecture, a client consists of any machine connected to a network that exchanges data with a server, and can also refer to the person using the machine.
clock cycle: Think of a clock cycle as one tick of the second hand (but generally at a much higher speed). Computer clocks run voltage through a tiny crystal that oscillates at a predictable speed to give a meaningful timing method to the computer. One clock cycle doesn't necessarily mean that the processor does one operation. Today's high-end processors often complete more than one operation per clock cycle, and sometimes, it will take several clock cycles to complete one operation.
clock speed: The clock circuit uses fixed vibrations generated from a quartz crystal to deliver a steady stream of pulses to the CPU. This refers to the speed in MHz of a microprocessor. It is one way of gauging the performance of a microprocessor. However, different processor architectures dictate that a higher clock speed in one architecture does not always mean better performance over a lower clock speed in another architecture.
back to top
DRAM (Dynamic Random Access Memory): This is the most common form of computer memory. It needs to be continually refreshed in order to properly hold data, thus the term "dynamic."
DSL (Digital Subscriber Line (DSL): This is a form of high-speed Internet access currently competing with cable modems. DSL works over standard phone lines.
DSP (Digital Signal Processor): A DSP is a microprocessor designed to work with analog signals such as video or audio that have been digitally encoded. The DSP then takes these digital representations and performs operations on them. DSPs are used in video, sound, and modem technology.
back to top
embedded processor: This refers to a microprocessor used in an embedded system. Typically these processors are smaller, consume less power, and utilize a surface mount form factor, as opposed to more standard consumer processors. Embedded processors are sold to consumers only in embedded systems.
embedded system: This refers to a system that is located entirely on a processor. All logic is contained in a single chip. New cars have many embedded systems working to keep emissions low and performance high.
Ethernet: This is a network topology that is able to send data at 10 Mbits/second. Workstations can exist on the same cable, but only one can communicate at a time. To get by these limitations, switched Ethernet and Fast Ethernet were invented, and can also be combined.
back to top
fabless: This term refers to a company that produces chips but doesn't own a fabrication plant, or fab. These companies are starting to become more and more successful at creating chips and renting out other companies' excess fabs to produce their chips.
fabrication plant (fab): A fab is a factory that takes raw silicon wafers and creates chips with them. Often, fabs are categorized by what micron process they use.
FDDI (Fiber Distributed Data Interface): This is a fiber optic interface that allows data to travel extreme distances (many miles/kilometers) without signal loss. It is far superior to copper wire for data integrity as well. FDDI is often used to connect Internet, LAN, or WAN backbones together due to these superior properties.
fiber optic: A method of transmission alternative to physical copper wire. The way it works is by pulsing light down a strand of glass. These pulses represent binary code. The advantage is that a single strand of fiber optic can carry thousands and thousands of different frequencies at once without data loss.
FIFO buffer: First In First Out Buffer buffer is an area of memory that holds information in the order in which it was received until the computer has time to use it.
firewall: This is a form of Internet security that stands between a private network and the Internet. It is like a wall in that it can prevent unwanted traffic from passing either way. True firewalls generally support packet-filtering, proprietary application filtering, and some proxy functions.
foundry: This is a synonym for fabrication plant. A foundry is the actual location that produces of microprocessors. The term foundry initially referred to a location where metal was poured into molds, but in technology speak, it refers to the place where microprocessors are molded.
FPGA: Field Programmable Gate Array is a microchip that can be made with thousands of programmable logic gates. Good features of FPGAs include short development times and low costs of production. FPGAs are often used for prototype or custom designs, including DSP and logic emulation.
frame relay: This is a packet-based communications method for connecting networks. Nowadays, it is commonly used to interconnect remote offices over the Internet or private LANs. Frame relay has no error checking, and assumes that devices on either side will be able to check for errors themselves. A frame relay connection can use an ISDN line for slow speeds, or can be over a T1 line or better if faster speeds are needed.
Function: In programming, a self-contained software routine that peforms a job for the program it is written in or for some other program. The function performs the operation and returns control to the instruction following the calling instruction or to the calling program. Programming languages provide a set of standard functions and may allow programmers to define others. For example, the C language is built entirely of functions.
back to top
Gbps (Gigabits per second): This is a rate of data transfer of 1 gigabit of data per second.
gigabit (Gb): This refers to approximately 1 billion bits.
gigabyte (Gig, GB): This refers to approximately 1 billion bytes.
GUI (Graphical User Interface): Graphical User Interface is any system that uses graphics to represent the functions of a program.
back to top
hash table: A table that contains "hash values," generally used for speeding up searching algorithms on sorted data. For example, a hash table for a dictionary might contain all the letters of the alphabet and the page numbers where each letter starts. That way, instead of searching through from start to finish for each word, you just use the hash table and get a big head start.
HDSL: High bit-rate DSL is a form of DSL that provides T1 (1.5Mbits/second) or better transfer rates and is meant to displace business T1 connections. HDSL requires much less bandwidth than standard T1 encoding to transfer the same amount of data.
heatsink: This is a device that contacts a microprocessor (or other object in need of cooling) and removes heat by exchanging it with air in a more efficient manner than the flat surface the heatsink is attached to. The heatsink does this by providing more surface area than a flat surface.
hertz: A measure of speed. One Hertz means one time per second, so one megahertz (MHz) means 1 million times per second. This is the common measure of speed for processors and electronic activities inside a computer.
hub: A central connection point. Standard terminology for a device that connects multiple computers in a network.
back to top
integrated circuit (IC): This refers to a combination of multiple circuits into a single integrated device. Today, the common microprocessor uses many millions of transistors, with each transistor counting as a single circuit.
Internet Protocol (IP): This is a connectionless communications protocol that forms part of the basis for the TCP/IP protocol suite. It is a fast protocol, but it has no mechanism for sequencing or error conditions. Error packets are simply lost.
intranet: Local network of computers using TCP/IP as the standard communications protocol. Usually an Intranet features some sort of HTML content that you can use a browser to look at. Think of it as a mini, private Internet
IP Address: The specific network address of a computer on a network using TCP/IP as its network protocol.
IPv4: Internet Protocol version 4 is an outdated version of the IP protocol that is still in use on the Internet. It uses a 32-bit addressing scheme, represented by 4 8-bit (0-255) numbers separated by periods, such as: 123.3.12.255 The addressing scheme allows for a maximum of about 4.3 billion numbers (256*256*256*256). This gets to be a problem as more and more devices are connected to the Internet.
IPv6: Internet Protocol version 6 is the current version of the IP protocol that features a 128-bit addressing scheme, as opposed to the 32-bit addressing scheme of IPv4, supporting a much higher number of addresses. It also features other improvements over IPv4, such as support for multicast and anycast addressing.
ISDN (Integrated Services Digital Network): ISDN is a digital line that is often used to connect to the Internet. It generally comes in two flavors: one is a 56 Kbps version, which in actuality only uses half of the ISDN line's bandwidth; the other is the 128 Kbps version, which uses both the 56 Kbps channels on the line. However, that's only 112 Kbps - the other 16 Kbps are an 8 Kbps back channel of each line.
ISO 9000: This is a certification granted by the International Standards Organization to companies that manufacture goods. It specifies that they have properly recorded all of their procedures and they meet with certain specifications.
back to top
JEDEC: Joint Electron Device Engineering Council is the standardization body of the Electronic Industries Alliance trade association. It was created in 1958 to form discrete semiconductor device standards, and was expanded to include integrated circuits in 1970. An example of a JEDEC standard is SDRAM memory.
jitter: This referes to delay in transmission of data that needs to be contiguous, like voice or full motion video. The information must be received in the same speed that it was recorded, not sent. It can be downloaded in stages but must be "reassembled" and finally viewed or heard at the speed that the spoken voice or visual must be experienced. It is a key variable in Quality of Service (QoS).
back to top
L1 cache: Level 1 cache is a small piece of very fast memory that's on the CPU chip itself, usually meant for holding instructions as they get executed. It sits between the CPU registers and the L2 cache.
LAN (Local Area Network): A small isolated network at one office location. Most office computers are connected to a LAN, but may also be connected to the Internet, or a WAN.
LANCAM: CAM working on a Local Area Network, A group of personal computers linked together to share programs, data, and peripherals.
latency: This is a measure of how long, on the average, it takes to get a response from something. This term usually refers to network response speed.
LSI (Large Scale Integration): This refers to chips containing thousands of transistors.
back to top
MAC (Media Access Control): Physical access card or interface device.
MAC address: This is a unique 48-bit address of a network card or device. The first part of the address is unique to the company that produced the device, and beyond that, it is a sequence of digits unique to a single device manufactured by a company.
mask: When comparing packet header infomration, portions of the search field may be eliminated because they may be the same for every address. This facilitates searching for a match to only the variable information. Masks can be redefined and reconfigured depending on the database being searched.
micron: This is the length of one millionth of a meter, or 1/1000000 meters. This length is also referred to as a micrometer.
micron process: This refers to the size of lines capable of being etched onto a chip. A fabrication plant with a .25 micron process can create smaller chips that consume less power, and can run faster than a fab with a .35 micron process.
microprocessor: (1) A central processing unit (CPU) fabricated on one or more chips, containing the basic arithmetic, logic, and control elements of a computer that are required for processing data; (2) An integrated circuit that accepts coded instructions, executes the instructions received, and delivers signals that describe its internal status. The instructions may be entered or stored internally. Also called "MPU" (microprocessor unit). Widely used as control devices for household appliances, business machines, toys, etc., as well as for microcomputers. Intersil is the industry leader in 16-bit CMOS microprocessors.
MIPS (Millions of Instructions Per Second): This is a rating of how quickly a processor can process instructions. A processor rated at 100 MIPS can process 100,000,000 instructions per second.
MPU (microprocessor): This term is commonly used to describe the CPU (Central Processing Unit.) More specifically, it refers to the part of the CPU that actually does the work, since many CPUs now contain L1 and L2 caches on-chip.
Multicast IP: This is a form of TCP/IP being proposed that will allow for high-bandwidth transmissions (like television channels) to be broadcast over the Internet to all the possible routers connected to end users. The benefit can be demonstrated with an example of 1000 people making separate connections. Each packet of data sent out would have to be sent to 1000 people. Thus, you have 1000 conversations active and with poor scalability. Multicast IP would have perhaps 10 routers with these 1000 people connected. Each packet would be sent instead to only 10 places, each end user's machine downloads packets directly from the router it is connected to.
multimedia: This refers to any use and transmission of audio and video data in a computer.
MUX (multiplexer): This is technically a logic circuit that sends one of several inputs out over a single output channel. In the network world, it is used to describe devices that send several signals over a single line at the same time, sequenced by time or frequency intervals. The device on the other side of the wire that receives the signal is a demultiplexer.
back to top
NIC (Network Interface Card): The NIC is an add-in board that enables a computer to connect to some form of computer network.
node: This refers to one computer / machine or address on a network. If you managed a network with 10 printers, 50 servers and 150 client machines, you could say you managed a network with 210 nodes.
Non-Volatile RAM: A small amount of RAM that stores information even after you turn off your computer. Used in modems (for storing your settings) and in hardware keys for protecting software.
back to top
OC12: Optical Carrier Level 12 (OC12) - This is a fiber optic connection that can handle 622 Mbps, often used for the 622 Mbps ATM protocol.
OC24: Optical Carrier Level 24 is a fiber optic connection that can handle 1.244 Gbps connection speeds.
OC3: Optical Carrier Level 3 is a fiber optic connection that can handle 155Mbps, often used for the 155 Mbps ATM protocol.
OC48: Optical Carrier Level 48 is a fiber optic connection that can handle 2.488 Gbps connection speeds.
OSI Model: The Open Systems Interconnection Model is a way of representing the complexities of computer networking in a seven layer model, ranging from the physical hardware of networking, all the way up to how application programs talk to the network. The seven layers are: physical, data link, network, transport, session, presentation and application. The seven-layer OSI model can be used to help diagnose network problems.
back to top
packet: A data transmission going from one network node to another. Packets often refer to complete units of information, multimedia or data, sent over computer networks. Packets contain header information, source and destination address information.
Packet Classification: The identifying of packets for quality of service (QoS). Packets can be classified by source and destination ports and address and protocol type.
packet filter: This is anything that filters out network traffic based on sender's address, receiver's address, and the type of protocol being sent. Some routers support packet filtering, all firewalls should. MUSIC's Ethernet Packet Filters perform wire-speed look-up fuctions for 10/100 Mbit Ethernet bridges and switches. These devices provide a glueless interface between most Ethernet MAC devices and a MUSIC LANCAM address table
packet filtering: Discarding unwanted network traffic based on its originating address or range of addresses or its type (e-mail, file transfer, etc.). Packet filtering is generally performed in a router.
Packet Header Look-up: Packet header field of an Internet data packets contains information on the destination address and other tag bits for QoS, priority etc. The look-up requires checking all these fields and sending the information contained in these fields to a router for routing the Packet appropriately.
parser: Strictly speaking, this is something that separates data into more easily understood chunks. More practically, a parser is the part of a compiler that goes through a program and cuts it into identifiable chunks before translation.
PEP (Packet Express Architecture): MUSIC Semiconductor's proprietary architecture that segregates packet processing devices into a physical hierarchy of dedicated functionalities. The broadest level provides packet header search, sorting and prioritization; the next, packet header manipulation and sequencing logic; the last, serves as a processor bus interface if a RISC processor is used. This architecture affords optimal performance at the RISC and network processor level by manipulating the packet stream into its most efficient configuration.
PGA (Pin Grid Array): This is a form of chip with a group of pins in concentric rectangles, designed to fit into a PGA slot.
pipeline: This refers to the technique of processing multiple parts of an instruction at the same time. Many processers have two or more instruction pipelines. As one instruction is executed, the next instruction is being decoded and the one after that is being fetched from memory. The pipeline is only as fast as its slowest member.
pipeline burst cache: This is a form of L2 cache that uses pipelining to speed data access by being able to feed data to the CPU at the same time it pulls data from memory. The burst mode also allows the cache to fetch additional data from main memory before the CPU requests it.
processor clock: The processor is regulated by a crystal that vibrates at a regular frequency when an electrical current is applied to it. Each vibration counts as one tick of the clock.
processor cycle: This is what occurs during one tick of the processor clock.
protocol: This is the behavior that computers must follow in order to understand one another. Think of it as a language. If two computers don't use the same network protocol, then they cannot communicate.
back to top
QoS (Quality of Service): This is an effort to provide different prioritization levels for different types of traffic over a network. Text or still images can be sent in bursts just as long as the whole file is eventually displayed as a whole. However, video and VoIP need to be received at the same rate they were recorded. Transmission of multimedia has necessitated more sophisticated packet prioritization, sorting, queuing and scheduling to improve QoS and insure the integrity of whatever is being sent.
queue: This is a data construct that is first-in, first-out (FIFO). Queues are used throughout the architecture of computers and are necessary in programming languages to accomplish certain tasks. Information is recieved in one order, but individual elements of that input may later be sorted according to type.
back to top
RAM (Random Access Memory): The place in your computer that programs reside when running. You can access any part of the memory, and it can easily be overwritten with new values.
resistor: This is a two terminal electronic component that resists an electrical current and turns the extra current into heat. Resistors are rated in ohms and can be used to regulate the voltage and current in a circuit. Basically, a resistor resists the flow of electricity. Resistors are used to lower voltages and currents so that components that require smaller voltage / current can function properly. If a power supply puts out 12 volts, and you have a chip that runs on three, you need a resistor to cut down the voltage.
RISC (Reduced Instruction Set Computing): RISC chips use a simpler instruction set to get their work done. This results in more instructions that need to be processed by the processor, but they are easier to process, so the chips can process more instructions per clock cycle.
RJ-11: This is the standard phone jack connector and supports 4 separate copper wires. This is commonly called a POTS line, for "plain old telephone system."
RJ-45: This is the standard Ethernet network connector, supporting 8 copper wires. You plug an RJ-45 cable into an RJ-45 port.
ROM (read only memory): Memory containing a program, data, or information about the device that has been programmed onto the chip at the factory.
router: Hardware and software that routes data between similar or dissimilar networks on the protocol level. Routers can support more than one protocol. Routers can actively direct messages between separate networks. They make decisions about which path the network traffic follows. Each network connected by the router has a unique identity with its own Internet address.
Routing: The forwarding of data packets in packet-switched networks, to the intended address.
Routing co-processors (RCP): Boost performance of high speed IP switches
back to top
scalability: Scalability is the ability to grow with your needs. Scalable software or hardware means that you only buy the parts you need, and that either have the ability to grow by adding on as you do. Applications or systems that are scalable are able to dynamically accommodate larger amounts of users without program modifications.
SCSI (Small Computer Systems Interface): A standard data pathway used mostly for hard drives and CD-ROM drives, but also a common interface for scanners and sometimes printers. Currently the fastest (and the most flexible) method of interfacing with hard drives. It comes in two major varieties: SCSI-2 or WIDE SCSI.
SDK (Software Developer's Kit): This is a programming tool that is tailored towards a particular purpose. The kit includes a compiler, linker, and an editor. Most hardware manufacturers have sdk's available that work specifically with their hardware to make writing compatible applications easier.
SDRAM (Synchronous Dynamic Random Access Memory): Like DRAM but arranged with a more efficient data pathway which allows for faster throughput.
SDSL (Synchronous Digital Subscriber Line): This is a form of digital subscriber line that has the same transmission speed upstream and downstream. It is most often used for business use. Contrast this with the consumer version, ADSL.
Search engines: ASSP device that searches a table
semiconductor: A semiconductor is a substance (usually Silicon doped with Germanium or Arsenic) that selectively conducts electricity. The selection usually occurs by running another current through a different axis. Microprocessors are made of semiconductive materials.
server: This refers to a machine whose sole purpose is to supply data, so that other machines can use it. See also client/server. This also describes any software process that runs on a server machine, and responds to client processes or programs locally, or across a network.
silicon: An element commonly associated with glass. Called silica when bonded with oxygen. Sand and quartz is a form of silica. Not to be confused with "silicone." The main component of computer chips.
SNMP (Simple Network Management Protocol): SNMP is a protocol used to manage network devices, usually hubs and routers. It operates over UDP, which is part of TCP/IP. Devices that support SNMP send messages to a management console. The management console then stores these messages in the MIB (Management Information Base). SNMP is often used to get traffic reports on hubs and check network utilization on various devices.
socket: A socket is composed of an IP address and a port number. For example, a socket could be referenced by "206.119.148.38 port 80." See also TCP/IP and ports. Another meaning to socket refers to the pin grid array interface between some CPUs and what they plug into.
SOHO (Small Office / Home Office): This refers to a class of equipment purchasers characterized by their requirements for low-cost but functional computers, fax machines, printers and other office equipment. Most computer makers target this market as a separate market segment between that of standard consumers and small businesses.
SONET (Synchronous Optical NETwork): This is a fiber optic network in a ring topology, often used to carry voice signals or Internet traffic. One ring is described as hot, or carrying data. The other ring is the protect, or standby, ring that transmits data in the opposite direction in case of service degradation or a fiber cut. Typically within 50 milliseconds of a failure, SONET equipment on the edge of the failing segment attempts to construct a new ring using a combination of standby fiber and hot fiber.
SQL (Structured Query Language): This is a type of relational database that features data in rows and columns. Each row represents a record. Each column is a field. Most serious database solutions today use some form of SQL to store and access large quantities of data quickly and reliably.
SRAM (Static Random Access Memory): Much more expensive and physically larger than DRAM, but much faster. The fastest SRAM has a latency of about 5 nanoseconds.
subnet: This is a term that refers to a group of TCP/IP addresses defined by a subnet mask. A subnet mask determines how many addresses are in the same subnet as a particular TCP/IP address. Addresses in your subnet are reachable without going through a router, and thus can be reached by broadcasts. To reach addresses outside of a particular subnet, you must transmit through a router. This is all part of the TCP/IP protocol.
Switches: An analog IC (typically CMOS) which, on command, either passes or blocks an electrical signal. Intersil is the leading worldwide supplier of DI (dielectric isolation) and JI (junction isolation) analog switches. See DI and JI.
synchronous: This is an outdated method of data transfer by which large chunks of information were sent as contiguous blocks. If transmission was interrupted, the entire block of data was lost. (See asynchronous). Synchronous now generally refers to a timed, or cyclic activity.
system bus: The path between a microprocessor and the chipset on the motherboard. This bus may or may not run at the same speed as the memory bus.
back to top
T1: An AT&T term for a digital carrier facility used to transmit a DS-1 formatted digital signal at 1.544 megabits per second. Also, a leased-line specification providing for 24 frames with an aggregate bandwidth of up to 1.54 Mbps.
T3: A term for a digital carrier facility used to transmit a DS-3 formatted signal at 44.746 megabits per second.
TCP (Transmission Control Protocol): This is the part of the TCP/IP suite of protocols that is responsible for forming data connections between nodes that are reliable, as opposed to IP, which is connectionless and unreliable.
TCP/IP (Transmission Control Protocol / Internet Protocol): The TCP/IP suite first saw usage on the original Department of Defense Internet in 1983. It’s first implementation was amazingly successful, and it is still THE protocol of the Internet. In fact, it has grown even more and is being used in private networks around the world. TCP/IP is a suite of communications protocols that allows communication between groups of dissimilar computer systems from a variety of vendors.
thermal compound: This is a paste that is applied between a microprocessor and heatsink. This paste fills in any tiny gaps that may be present on the visibly flat surfaces of the microprocessor package and heatsink. Without thermal compound, cooling may be less efficient as the heatsink and microprocessor package may have air gaps between them. Thermal compound has much more efficient heat exchange properties than air.
Token Ring: This is a network topology pioneered by IBM, and eventually made into the IEEE 802.5 standard. The original version transmitted data at 4Mbits/second, and it was updated to transmit at 16Mbit/second. Token ring networks are wired in a ring topology, and nodes on the network pass a token around. Whichever node has the token is allowed to use the network.
transistor: An electronic device that acts like an electrically activated switch but has no moving parts so it can switch millions of times per second.
TTL (Time to live): This is the maximum amount of time that a data packet can travel on a network before being discarded and assumed to be lost. Also, Transistor Transistor Logic, which is a specific method of wiring a digital circuit using bipolar transistors.
back to top
URL (Universal Resource Locator): This is what is used to give Web addresses for HTML, VRML, WAV and other files. It simply contains the Internet name of the machine containing the data and the path to the file. The address also includes what protocol should be used.
back to top
virtual memory: This is a section of a hard drive used to augment a computer’s RAM. Different operating systems have different ways of dealing with virtual memory. UNIX and Linux use a separate hard drive partition dedicated to virtual memory, while Windows can use part of any hard drive partition. Virtual memory is used when the operating system runs low on physical memory or RAM. Virtual memory is also used to swap out lesser used portions of physical memory, freeing it up for other operations.
VLAN (Virtual LAN): A VLAN is created when a bunch of physically connected ports are grouped together by network hardware that supports VLANs. These VLANs are each treated as completely separate entities and can only be joined together by a router. This scheme is useful for grouping departments together for security and minimizing network traffic.
VLSI (Very Large Scale Integration): This refers to the amount of transistors that are incorporated in a chip. A VLSI processor has on the order of 100,000 or more transistors.
VMEbus: This is a backplane interconnection bus standard developed by Motorola and others and now and IEEE standard. It has data bus sizes of 16, 32 or 64 bits and VMEbus boards can hold CPUs or peripherals.
VoIP (Voice over IP): This is the practice of using an Internet connection to pass voice data using IP instead of using the standard public switched telephone network. This allows a remote worker to function as if they are directly connected to a PBX even while at home or in a remote office. As well, it skips standard long distance charges, as the only connection is through an ISP. VoIP is being used more and more to keep corporate telephone costs down.
back to top
wafer: This refers to a flat round piece of silicon that is used in the manufacture of microprocessors. Fabrication plants, or fabs, typically will take a wafer and carve many several microprocessors into it. Over the years, wafer size is increasing, so that more chips can be fit on a single wafer. Wafers are created by taking a cylinder of silicon and slicing it - much like taking a loaf of bread, and cutting slices - but to much more precise measurements.
WAP (Wireless Application Protocol): This is a proposed standard that allows for transfer of data securely between wireless devices, such as PDAs, cell phones, pagers or other combinations of those devices. WAP supports many different wireless networks.
word: This is a group of bits of data regarded as a whole while programming or transferring data. Often a word is 8 bits in length, also referred to as a byte.
word size: This term refers to the number of bits of data stored in a CPU or memory register. Typically, the number is a power of 2, with 8, 16, 32, and 64 being common. You have to deal with word size when doing certain data manipulations while programming.
back to top
X.25: This is packet-switching service that connects remote terminals to host systems. X.25 has higher overhead than Frame Relay, but has been around longer and is better supported. X.25 predates the OSI model.
back to top
zip: This is a file extension added onto files created by the PKZip file compression program written by PKWare. Zip files are the most common compressed file format in existence and are supported by most compression and decompression programs such as WinZip and WinRar.
back to top