IP telephony

from Wikipedia, the free encyclopedia

IP telephony (short for Internet Protocol - telephony and Internet telephony ) or Voice over IP (short VoIP ; from English voice over internet protocol , for literally vocal [transmission] on [the] Internet Protocol ), is making phone calls over computer networks , which are built according to internet standards. Information typical of telephony, including language and control information, for example for setting up a connection , is transmitted over a data network. Computers , telephones specializing in IP telephony, or classic telephones that are connected via special adapters can establish the connection with the call participants .

IP telephony is a technology that enables telephone service to be implemented on an IP infrastructure, so that it can replace conventional telephone technology including ISDN and all components.

The aim of using IP telephony by communication network operators is to reduce costs through a uniformly structured and operated network. Due to the long service life of classic telephony systems and the necessary new investments for IP telephony, the change with existing providers is often implemented as a long, smooth transition ("smooth migration"). Meanwhile, both techniques exist in parallel. This results in a clear need for solutions to connect both telephony systems (e.g. via VoIP gateways ) and the need for targeted planning of the system change, taking into account the respective options for cost and performance optimization. The number of providers exclusively with new technology (i.e. IP telephony instead of conventional telephones) is increasing. At the end of 2016, around 25.2 million people in Germany were using Voice-over-IP technology.

Switching VoIP calls - switching service

The switching of telephone calls is an essential task in computer networks. Since many users are dynamically connected to the Internet, so that the IP address changes frequently, the IP address itself cannot be used as a "telephone number" for contacting the VoIP telephones. A switching service in the form of a server takes on this task and enables telephony when the IP addresses of the IP telephones change.

  • VoIP telephones register with the server (for example SIP server); therefore the server knows the current IP address of the phones.
  • With the help of the IP address of the phone, which has been made known to the server, it can take over the switching, and the selected IP phone rings depending on this IP address (i.e. anywhere in the world, if the IP -Telephone registered from there with the switching server via the Internet).
  • Communication between the IP telephones can take place independently of the server.
  • There are commercial services that offer a local telephone with an account for the switching server, which can also be reached via the landline network. The IP calls are usually free.
  • If there is a fixed IP address, it is possible to operate a mediation server on the associated computer ( e.g. OpenSIPS ) in order to connect multiple mediation servers to one another , comparable to the connection of several local area networks in the fixed network . In commercial solutions, there are often partner networks that create a free connection between VoIP partner networks. The network selection is often limited, as the companies have to generate their sales with the connections from VoIP telephones to the regular landline network . From a technical point of view, free, self-operated open source telephony servers can form a network of exchanges on the Internet, regardless of these economic limits. Even if SIP telephony servers work technically well, there is currently no institutionalized networking of such SIP switching servers.


In addition to the telephone networks, a further communication infrastructure gradually emerged on the lines of the telephone networks . Beginning with the networking of EDP systems in the 1980s, followed by the Internet development in the 1990s, the transmission performance rose and rises continuously: While 300 bits per second were initially achieved with acoustic couplers , up to 100,000,000 in January 2008 Bit per second for end consumers with DSL connection on normal house phone connections or in the cable network. This infrastructure forms a basis for IP-based data networks, especially for the Internet as a public network.

In 1973, the first digital voice transmissions were implemented in the Arpanet using the Network Voice Protocol between PDP-11 computers. A data transmission rate of 3490 bit / s was made available to the voice channel . Only four years later, the Network Voice Protocol described above went into the RFC 741 standard , before the Internet Protocol (IP) was specified in RFC 791 in 1980 . Also in 1980, the first recommendations of the ITU-T (at that time still CCITT ) for ISDN were documented, which was introduced commercially from 1989 and enables calls with higher voice quality and additionally integrates various services such as telephone number transmission in a network. The standard data transmission rate of ISDN grew from 3490 bit / s with NVP-II to 64 kbit / s. In the same year the development of the World Wide Web began , which would later prove to be the basis for the broad success of the Internet.

With the GSM mobile radio, a service for mobile voice transmission with a data transmission rate of 13 kbit / s (260 bit frames with a frame duration of 20 ms) was created in Germany ( D-Netz ) in 1992 . However, these 13 kbit / s only relate to the transmission rate of the user data . To protect the transmission of user data against transmission errors, e.g. To protect against e.g. atmospheric disturbances, redundancy is added to the signal by the channel coding . This allows the data frame to grow from 260 bits to 456 bits, while the frame duration must remain constant because of the real-time requirements for voice connections. The gross bit rate of the transmission (user data + redundancy for error correction) is therefore 22.8 kbit / s.

In 1994 Michaela Merz developed mtalk with the Free Software Association of Germany, a free Voice-Over-IP software for GNU / Linux and Unix. The first versions of mtalk only had rudimentary data compression . mtalk formed the basis for a whole range of VoIP software. For historical reasons, various packets are still available for retrieval from various servers.

In 1995, a Windows program by the Israeli company Vocaltec Communications enabled Internet telephony, but only in half-duplex operation , which is why the conversation partners could only speak alternately with poor voice quality. Connections to computers that did not use the same software were not possible. Just one year later, QuickTime conferencing enabled audio and video communication in full duplex mode via AppleTalk and IP networks on the one hand, and the Real-Time Transport Protocol was described in RFC 1889 on the other .

Three years later, in 1998, an ITU-T framework standard was adopted for the first time with H.323 , so that solutions from different manufacturers should be compatible with one another. The Session Initiation Protocol (SIP) in RFC 2543 was specified in the following year. The subsequent establishment of VoIP solutions in 2001 in Austria resulted in the first notification by the regulatory authority to IPAustria of the operation of a carrier voice switching network . In the sense of today's VoIP, the SIP extension in RFC 3261 followed in 2002 to improve VoIP , as well as the adoption of ITU Q.1912.5 for interoperability between SIP and ISDN user part for better connection to other networks .

Contrary to the standardization of VoIP, the software Skype appeared in 2003 , which uses its own, non-disclosed protocol for IP telephony based on peer-to-peer technology.

Working principle

VoIP encapsulation

Telephoning with IP telephony can present itself to the subscriber in the same way as in classic telephony . As with traditional telephony, the telephone conversation divides it into three basic operations, the connection, the call transfer and connection release . In contrast to traditional telephony, VoIP does not use dedicated “lines”, but rather the voice is digitized and transported in small data packets using the Internet protocol.

Signaling protocols

The establishment and termination of connections (connection control, signaling ) is carried out using a protocol that is separate from voice communication. The negotiation and exchange of parameters for the voice transmission take place via different protocols than those of the connection control.

In order for a connection to a call partner to be established in an IP-based network, the current IP address of the called subscriber must be known within the network, but not necessarily on the caller's side . Geographically fixed connections as in the fixed network (PSTN) do not exist in purely IP-based networks. The accessibility of the called party is made possible, similar to in cellular networks , by prior authentication of the called party and the associated announcement of their current IP address. In particular, a connection can thereby be used independently of the user's location, which is referred to as nomadic use .

A fixed assignment of telephone numbers to IP addresses is not possible due to a change of location of the participant, change of user on the same PC or dynamic address assignment when establishing a network connection. The solution generally used is that the participants or their end devices store their current IP address on a service computer ( server ) under a user name. The computer for the connection control, or sometimes the end device of the caller itself, can request the current IP address of the desired call partner from this server using the selected user name and thus establish the connection.

Common signaling protocols are:

Establishing a connection with SIP

The Session Initiation Protocol (SIP) was developed by the Internet Engineering Task Force (IETF). Like H.323 , the manufacturer-independent specification of SIP enables the use of SIP-based systems in heterogeneous environments, in particular the coupling of VoIP components from different manufacturers. As with other standards, the interoperability of components is not guaranteed by compliance with the specification (SIP compatibility) alone, but must be checked in individual cases by means of interoperability tests. Basically, SIP is suitable for application scenarios beyond VoIP and video telephony.

With SIP, the participants have a SIP address (similar to an e-mail address) in Uniform Resource Identifier format (URI format), such as " sip: 0123456789@example.com" , where "0123456789" denotes Username and "example.com" represents the domain. SIP end devices ( user agents ) must register the IP address and port at which they can currently be reached via SIP with the SIP registrar server of their domain. By default, the Domain Name System (DNS) provides information about the responsible SIP server for a domain. Procedure for establishing a connection:

  1. The end device of the caller sends a message with the SIP address of the called party to the server of its own SIP service provider (SIP proxy).
  2. This connection request is forwarded by the SIP proxy of the own domain to the SIP proxy of the called domain. With the help of the SIP location service, this determines the IP number and port of the called SIP address and forwards the message to the end device of the called party.
  3. If the connection request can be processed there, the terminal sends a corresponding message back to the caller via the server.
  4. At this point, the called party's terminal rings and the caller hears a ringing tone .

As part of the establishment of a "session", all relevant information about properties and capabilities is exchanged between the end devices. A direct communication between the two terminals has not yet taken place. The servers are no longer necessary for the actual telephone call, the end devices send their data directly to each other and the data exchange during the call bypasses the server. The Real-Time Transport Protocol (RTP) is usually used to transmit this data in real time .

To end the call, one of the terminals sends a SIP message to the server, which forwards it to the other participant. Both terminals terminate the connection.

Like H.323, SIP provides the option of establishing a direct connection between two end devices without the use of SIP proxy servers, only via the IP address. To do this, however, all existing entries for SIP registrar servers must first be deleted on many terminals.

Number systems

Although the IP addresses of the participants can be used to establish the connection, these are not always known to the users and can change. There are therefore a number of approaches to give the subscribers an individual and mnemonic inexpensive Internet telephone number that is independent of the IP address. Starting with pure SIP numbers, there are approaches for integrating Internet telephony into the existing numbering plan of conventional telephone networks through to a completely new system. Important aspects of the European Union and the German Federal Network Agency (BNetzA, formerly: RegTP) are compliance with regulations and, in the medium term, the integration of emergency call systems.

SIP address

Many service providers offer SIP addresses for users who want to make phone calls to other Internet users over the Internet. Unlike telephone numbers or MSNs , SIP addresses are not tied to a connection, but can be used from any Internet connection in the world like e-mail accounts. Although this applies to telephone numbers that are assigned to a SIP address for incoming connections, the SIP address offers advantages above all for the caller. For example, telephone connections using the SIP address are possible between two terminals instead of always having to be routed via the telephone network, as is the case when dialing a telephone number.

To get your own SIP address in URI format, you have to register with one of many free or paid providers. Since many providers either only assign SIP addresses with pure sequences of numbers (e.g. 12345@example.com) or assign a numerical alias to the non-numerical address, IP telephones with a normal keypad can be used to dial in order to dial parties who are using the same SIP -Server registered. Customers of a SIP service provider can be dialed via their SIP address and call others, provided the provider of the called party allows the external SIP request. Most providers of SIP addresses allow access from the conventional telephone network, as they can generate additional income through the termination fees (the transmission from the telephone network to the connection of the called subscriber). Via this detour, which is subject to a fee, the subscriber can call other SIP service providers if their own provider or that of the other party blocks accordingly. There are agreements between some providers that allow customers to communicate directly with one another via telephone number. In this case, an Internet connection is established between the participants, but with the participation of both SIP providers. It is usually possible to dial the "internal telephone number" (that is the part of the SIP address before the @ sign) within the same provider network using a standard telephone with a number field. For this reason, most SIP addresses contain only digits in this part.

Many SIP adapters that are designed for connecting a conventional telephone with a number pad offer the option of saving SIP addresses in the internal telephone book instead of a telephone number and triggering this SIP address using an assigned speed dial on the telephone. In these cases, SIP addresses can be dialed at least indirectly using a conventional telephone.

Phone number

A telephone number is not absolutely necessary for IP telephony. However, since most connections are made using the conventional telephone network, a SIP address must be assigned to a conventional telephone number at least for incoming connections. A telephone number is not required for outgoing calls. To transmit a valid telephone number as sender identification, in addition to the "internal telephone number" (see SIP address), many providers can use the CLIP (no screening) function, which transmits a user-defined telephone number via which the user can be reached in detail. In some countries (including Germany) it is required that the provider verifies the specified phone number as belonging to the customer via a callback (e.g. teledialog system with PIN transmission).

The separation between providers for incoming and outgoing connections makes sense if the internet service provider already has a telephone number for incoming connections and an alternative (often cheaper) provider is only required for outgoing connections. For this reason, most free providers only offer a telephone number as an option for a surcharge, especially if a free telephone connection without a flat rate is offered.

There are basically two options for switching a telephone number:

  • Most Internet telephony providers offer phone numbers for incoming calls, as this enables them to generate additional income.
  • Other providers - such as the services of the Dellmont Group (Voipbuster, Megavoip, etc.) - offer the option of mapping (assigning) the DDI number (Direct Dialing In) registered with a third-party provider to your own SIP connection. In this case, there is no need to port the number when changing the SIP provider. This option of having the telephone number and SIP account managed by separate providers has not yet become generally accepted in Germany, but is quite common in other countries.

Some providers do without the switching of incoming calls and do not offer this option either.


Telephone numbers can be looked up on the Internet using Telephone Number Mapping (ENUM). This process is being promoted by some network operators and by both the German ( DENIC ) and the Austrian ( Nic.at ) domain registry.

With ENUM, the number is reversed and provided with dots between the individual digits, as a subdomain of the top-level domain "arpa" with the second level domain "e164" in front. For example, +49 12345 6789 becomes This solution assumes that the telephone customer already has a phone number.

Due to the EU directives on number portability when changing the telephone provider , ENUM (at least in Austria) is experiencing the hoped-for upswing. Before telephone providers arrange a telephone call based on their own databases, they check whether there is a DNS entry for the number called and the service used at ENUM . If so, the call is switched to the address specified in the DNS ( PSTN or SIP subscriber).

ENUM's public approach is unpopular with large commercial providers. On the one hand, it enables attackers to use automated free advertising calls, so-called SPIT ( Spam over IP Telephony). On the other hand, customer data could be requested. With suitable measures, ENUM directory operators can prevent automated mass queries so that both dangers can be limited. Another, perhaps essential, reason why many providers are reserved with regard to ENUM is that there are no sources of income due to free calls.

Conventional local numbers via a gateway

VoIP providers can use their own gateways to obtain free telephone numbers from the number supply of the German local networks and assign them to their customers. These customers can also be reached from the conventional telephone network. However, the Federal Network Agency limits such offers to participants who have their place of residence in these local networks. The reason, which is difficult to understand for a location and connection-independent service, is that otherwise the reference that the area code has to the place of residence will be dissolved. The providers are therefore obliged to check whether the customer actually lives in the desired local area network and to purchase numbers from all local area networks in which they (want) customers. For cost reasons, most of the smaller VoIP providers only offer numbers in the larger local area networks. If the customer lives outside an available area code, many providers provide 0180x numbers. However, this procedure is only permitted on a transitional basis.

If the VoIP provider uses the SIP protocol when establishing the connection, the customer has a SIP number in addition to the local number. However, many providers only give their customers the fixed network phone number assigned. In addition, many of these providers block Internet calls from callers who have not registered with them or one of their partners. This means that an Internet telephone call can only be made free of charge if both parties to the call have registered with the same provider (or a partner provider).

For most companies and authorities, adopting the entire existing numbering plan of the existing conventional connection ( area code , main number and all direct dialing numbers ) is a prerequisite for switching to an IP telephony service provider. Only a few providers offer this for SIP.

Special internet numbers

In Austria, the area code +43 780 and the location-independent area code +43 720 were created especially for convergent services - which includes Internet telephony. A similar solution was recommended by the German regulatory authority. After entering a 032 area code , a VoIP operator can be selected - similar to mobile communications with a "block identifier" - in order to then dial the actual end number of the subscriber. The 032 subscriber number is assigned independently of the local network boundaries of the geographical telephone number and can therefore be retained when moving to other local networks. Since there is no explicit geographic location associated with the 032 area code, the 032 numbers are generally predestined for nomadic use at different locations.

In the past, the 032 numbers could not prevail with most VoIP providers, but were used, for example, by the two largest national telephone companies ( Deutsche Telekom and Vodafone (formerly Arcor )) for their VoIP offers and increasingly for other value-added services. In the meantime, only a few call-by-call providers are unable to reach the number range 032 ; The numbers can be reached from the mobile phone networks since the activation by the last missing large mobile phone operator , Vodafone, in October 2007.

Often the costs for calls to 032 numbers from the cellular networks are significantly higher for customers than for calls to the fixed network. Calls from the landline network to a 032 number, on the other hand, are often treated the same as normal telephone calls in terms of charges, but not included in existing flat rates; so z. B. with telecom connections.

Call transfer

Principle of a conversation via IP telephony with the possible use of an IP telephone

As in traditional telephony, the acoustic signals of speech are first analog with a microphone converted (via the handset) into electrical signals. These analog electrical signals are digitized ( encoded ). They can optionally be compressed (ITU-T G.723.1 or G. 729 Annex A are widely used for this) in order to reduce the amount of data to be transmitted. The data converted in this way is transported via a public or private telecommunications network. Due to the packet switching method used for transport , the data is divided into many small packets.

Digitization of the analog signals and digital processing

The analog speech signal is sampled for digitization at a suitable sampling rate and the results (samples) are converted into a regular sequence of digital signals by an analog-to-digital converter (ADC).

The data rate of this digital data stream is the product of the sampling rate and the resolution of the ADC in bits. If necessary, it can be reduced by means of coding before transmission. Different compression factors are possible depending on the codec (coder-decoder) used. Many codecs use lossy processes in which information that is unimportant for the human ear is omitted. This reduces the amount of data and reduced as the required for transmission bandwidth significantly without the hearing impression to deteriorate significantly. If too much information is left out, there will be a noticeable deterioration in speech quality .

Different codecs that use different coding methods are used. Some are specially designed to achieve a data rate that is significantly lower than the 64 kBit / s of the ITU standard G.711 based on the standard telephone quality (sampling rate 8 kHz, 8 bit ADC resolution) . Other codecs such as G.722 (see HD telephony ), on the other hand, code on the basis of more highly sampled and resolved digital speech with radio or CD quality (7 kHz and more bandwidth of the transmitted speech) with a moderate need for transmission bit rates.

Depending on the digitization and coding process, the frequency range of the coded speech, the bandwidth required for transmission and the resulting speech quality (source coding) vary. In addition, the coding method can be designed in such a way that certain typical disturbances on the transport route are compensated for (channel coding). So that the data can be converted back into language that the human ear can understand after transport, the recipient must use a decoder that matches the coder, which means that many end devices contain several codecs to ensure interoperability.

Transport of the data

Normally, each end device sends the coded voice data "directly" via the network to the IP address of the remote station, regardless of the signaling. The call data does not flow through the server of a VoIP provider.

The actual transport of the data takes place via the Real-Time Transport Protocol (RTP) or SRTP and is controlled by the RealTime Control Protocol (RTCP). RTP uses the User Datagram Protocol (UDP) for transmission . UDP is used because it is a minimal, connectionless network protocol that, unlike the Transmission Control Protocol (TCP), was not designed for reliability. This means that the receipt of the voice packets is not confirmed, so there is no transmission guarantee. The advantage of UDP is its lower latency compared to TCP, as there is no waiting for a confirmation and incorrect packets are not sent again and the overall data flow is not delayed. Completely error-free transmission is not necessary due to the redundancy of spoken language (and the ability of the codecs used to correct errors). A short running time is much more important for a smooth conversation .

Transmission quality

The network requirements for data transmission and IP telephony differ considerably. In addition to the required transmission capacity (around 100–120 kbit / s for a call coded with G.711), quality features such as average delay , fluctuations in delay ( jitter ) and packet loss rate have a significant influence on the resulting voice quality. Through prioritization and suitable network planning, it is possible to achieve a voice quality and reliability comparable to conventional telephony, regardless of the traffic load.

Since the Internet in its current form (as of 2008) does not guarantee a secure transmission quality between subscribers, there may well be transmission disruptions, echoes, dropouts or connection interruptions, so that the voice quality does not come close to that of conventional telephone networks. In most cases, however, it is better than cellular telephony. With a good DSL connection (the bit rate in the direction of the network is the bottleneck , it should be between 120 and 200 kbit / s per telephone connection), the voice quality of a traditional telephone connection can be achieved at significantly lower costs.

QoS on Layer 3 with VoIP

It makes sense to identify and prioritize “voice packets” over other data packets on the Internet. The IPv4 protocol , which is still predominantly used in the Internet today, offers such possibilities ( DiffServ ), but they are not or not consistently observed by the routers in the Internet. Carefully planned and configured private IP networks can, however, guarantee an excellent " Quality of Service (QoS)" (also with Ethernet as the bit transmission layer) and thus enable telephony with the usual quality in the event of overload in the data area. However, the status quo on the Internet so far has been best-effort transport, i.e. the equal treatment of all packages. The mostly usable telephony quality is due to the overcapacities of the networks. A number of committees and research projects (MUSE, DSL Forum , ITU-T ) are working on further QoS standards for the future, multimedia- heavy Internet .

No significant improvements in QoS are to be expected from the follow-up protocol IPv6 . IPv6 brings a new element flows . So far there is still no clarity on how this should be used. Whether or not the infrastructure takes these markings (priority, DSCP code) into account is ultimately a financial question. The future will show whether Internet service providers will provide higher quality IP streams for more money.

Quality features

In order to be able to conduct high-quality communication via Voice-over-IP, the data packets used for voice transport must arrive at the counterpart in such a way that they can be put together to form a true copy of the original, temporally contiguous data stream. The factors listed below determine the quality of the system.

In the intranet , the network operator can autonomously determine the quality of voice transmission through the server configuration and router equipment as well as the distribution of the access points . In the Internet, the providers temporarily involved in the entire chain determine the transmission quality.


The required throughput (amount of data that can be processed by a system or subsystem per unit of time) depends primarily on the coding used. An uncompressed call typically has a data rate of 64 kbit / s (payload). Depending on the compression method used, the bandwidth required for pure IP telephony is a maximum of just under 100 kbit / s (64 kbit / s net plus the overheads of the various communication protocols).

Since the network is used jointly with other data services, a data connection (such as a DSL connection) with a bandwidth of at least 100 kbit / s in both directions is recommended , especially in the home . It should be noted here that in the frequently used ADSL process, the upstream bit rate is significantly lower than the downstream bit rate.

Running time (latency) and jitter

The transport of data takes time. It is (as a running time or latency English delay, latency ) and is in conventional telephony substantially the sum of the signal delay times of the transmission channels. In the case of telephony over IP networks, there are further delays due to the packaging and intermediate storage and, if necessary, data reduction, compression and decompression of the data. In telephony (regardless of the technology with which it is implemented), according to ITU-T recommendation G.114, up to 400 milliseconds of one-way transit time (mouth to ear) is the limit up to which the quality of real-time communication is still acceptable applies. However, from around 125 milliseconds, the runtime can be perceived as annoying by humans. The ITU-T therefore recommends generally not to exceed a one-way transit time of 150 milliseconds for highly interactive forms of communication.

The time fluctuation between the receipt of two data packets is referred to as jitter . To compensate for this, so-called “buffer memories” ( jitter buffers ) are used, which cause an additional deliberate delay in the received data in order to then output the data isochronously . Packets that arrive later can no longer be incorporated into the output data stream. The size of the buffer memory (in milliseconds) is added at runtime. It allows you to choose between more delay or a higher packet loss rate.

Packet loss

One speaks of packet loss when data packets sent do not reach the recipient and are therefore discarded. In real-time applications, the term packet loss occurs when the packet reaches the recipient but arrives too late to be able to be added to the output stream. Telephony is an ITU-T G.114 packet loss rate (packet loss rate) still classified as acceptable to a maximum of 5%.


The availability of the overall system results from the individual availability of the components involved and their interconnection (cascaded - in series, or redundant - in parallel). The availability of an IP telephony system therefore depends primarily on the network design. A US study from June 2005 examined the availability of IP telephony in the US. On average, almost 97% was achieved. This corresponds to a failure on a total of 11 complete days per year. In addition, many German DSL providers have a so-called 24-hour forced disconnection , which means that the line is disconnected when the line is constantly in use.


There are different architectures for VoIP. The following are widespread: the architecture according to the H.323 framework standard of the ITU-T , which provides for the elements terminal, gateway, gatekeeper and MCU, and the architecture according to the de facto standard SIP of the IETF . There are also a number of non-standard solutions for VoIP.


In ITU terminology, a terminal is the “multimedia end point” of communication, in the narrower sense the terminal device for inputting and outputting voice information. Its (approximate) equivalent in the IETF's SIP terminology is the user agent.

Terminal types

Terminal types

There are three basic types of devices that can be used for IP telephony.

  • With software running on the PC, a so-called softphone .
1140E VoIP phone
  • With an (S) IP telephone that can be connected directly to the local data network ( LAN ) or a WLAN telephone for wireless networks. In this case, a PC is not required for telephoning (except possibly for configuration work or to facilitate certain processes such as entering speed dials, entering alphanumeric data, etc.).
  • With a conventional telephone that is connected to the LAN via an analog or ISDN telephone adapter for VoIP ( ATA and ITA). ATA and ITA are offered directly as a connection option for telephones integrated in DSL routers. In this case, too, no PC is required for telephony operation, but it is necessary to set up the user data once. Terminals for GSM -Mobiltelefonie have the opportunity to lead IP calls in the available wireless network (see open source operating system OpenMoko ). For reasons of cost, these terminal types combine GSM mobile and IP telephony by using the cheaper IP telephony with the mobile phone when WLAN is available .

However, problems with the use of Voice over WLAN have so far been the lack of standards for bandwidth management over the air (too much user activity at the same access point causes a critical rate of packet loss in the VoIP connection) and for handover (connection interruption when the device moves to another access point ) as well as the high power consumption of battery-operated devices.

Fax over IP (Fax over IP, FoIP)

The T.30 protocol is used in the voice channel to send faxes via ISDN or analog connections . The high reliability of a voice channel connection in conventional TDM-based networks usually ensures secure transmission. However, this does not apply in IP networks, because speech is usually transmitted unsecured ( RTP over UDP ), despite the same coding of the language, such as the G.711 codec , which is used in TDM-based networks and IP networks. IP packets can be lost and up to 5% of losses are imperceptible to the human ear. The fax transport over an IP network by means of such a voice codec, a coding used for this which is optimized for human speech, leads, however, to loss of information or connection interruptions of the fax.

In order to be able to send faxes over IP networks, the following encodings and protocols are used in the voice channel:

  • Via a voice codec: Fax over VoIP, reliable transmission is not always possible
  • e-mail
  • T.37 (email based)
  • Real time: T.38

This results in different approaches to using fax over IP (FoIP).

  • A conventional analog fax machine is used in an IP network such as a TDM-based telephone network with an analog or ISDN connection. (This is the most frequently requested solution.)
  • A fax machine with direct T.38 or e-mail support and network connection while one available gateway with T.38 or e-mail support with access to the PSTN -Telefonnetz and a gatekeeper is used.
  • There are fax machines that are designed for direct fax transmission and reception via T.38.


The gateway is an intermediary between the two technologies.

In order to establish connections to conventional telephone networks, switching computers, the so-called gateways , are required. These are connected both to the communication network of the IP telephone and to the conventional telephone network ( PSTN ). If you receive a request from an IP telephone, you forward it to the telephone network by calling the desired number. If you receive a call from the telephone network, you forward a request to the corresponding IP telephone.


A gatekeeper is an optional component in the H.323 environment and fulfills central functions such as terminal registration or the establishment and termination of connections between registered terminals.

Multipoint Control Unit (MCU)

The optional Multipoint Control Unit (MCU) is used with the H.323 where connections between more than two terminals are required ( telephone or video conference ). This is where the terminal properties are negotiated and the conference is controlled. Possibly. different codecs and bit rates are implemented and the mixed information is distributed via multicast .

Areas of application

Direct internet telephony

IP telephony is used to conduct calls around the world directly over the Internet, known as Internet telephony . The classic telephone network is no longer used.

For end customers (private users and home offices), reasons for use are in particular:

  • Save fees with IP telephony. Analog or ISDN end devices, sound-enabled computers (preferably with handset or headset ) and special IP telephones can be used as end devices via special adapters ( ATA , ITA ) . There are no call charges for calls between two IP telephony subscribers.
  • The connection to and from participants in the conventional telephone network is possible. It is established through a gateway provided by the provider, the gateway service. Calls originating through gateways usually incur special charges.
  • Regardless of where you are, you can always be reached at the same address and phone number.

Organization-internal telephony

Within organizations such as businesses, IP telephony is increasingly used to bring the telephone network and the computer network together. The data transport of the telephone calls for the signaling and the transmission of the digitized speech takes place via the IT network ( LAN ). In this way, the infrastructure costs can be reduced through uniformity of cabling and active system components. The IP telephones are connected to the network connection like a workstation PC. Conventional end devices have to be replaced or adapted.

Telephony services, in particular subscriber management and call switching, are provided via IP-enabled telephone systems that are also connected to the network. Telephone systems at different locations can be linked to capacity reserves via the extranet ( WAN ) and existing data lines. Not all of these different locations have to be equipped with their own telephone system. Locations that do not have a local telephone system installed are called remote units. For connections to the conventional telephone network, such as the public telephone network (PSTN), so-called gateways are used between the IP network and the conventional network.

The structure of the overall system is described in so-called scenarios, which can contain several transitions between conventional telephony and VoIP. The changeover from traditional telephony to VoIP, known as migration , usually takes place gradually. Parts of a company, preferably new departments, are gradually being equipped with the new technology.

With combined telecommunications systems that provide IP and conventional ports , creeping migration (soft migration) is possible, as conventional connections can continue to be operated and gradually replaced by IP connections. These PBXs are called hybrid systems.

After a switch to VoIP, the voice quality and the reliability of the telephone technology depend entirely on the network technology , which must be taken into account when planning and administering the networks and places much higher demands on the hardware.

A cloud telephone system is a telephone system for companies that uses IP telephony and is not operated locally in the company, but on the outsourced servers of a cloud telephony provider . A cloud telephone system no longer requires a conventional telephone connection, but only requires an Internet connection and a VoIP device or softphone on a PC or mobile phone to handle calls .

Background technology of conventional telephony

Conventional telephone networks in Europe are based on the circuit-switched PCM30 method. The operators of telephone networks can use IP telephony for the transmission of calls without changing the call participants. IP telephony can be used for parts of the network or for the entire network.

Call-by-call providers, for example, have been using IP telephony for international connections for a long time . The calls are routed between the local telephone network and the telephone network of the destination country via the Internet, which results in cost advantages.

Next Generation Networks (NGN) only use packet switching networks for telecommunications. The aim is to use network resources more efficiently and to create a common platform for all services. There is a separation between the transport and service levels.

Connection prices

If both participants are connected to the Internet, there are normally no additional costs for Internet telephony apart from the costs for Internet use. In this case, calls using an open SIP server are free of charge worldwide for subscribers with an Internet flat rate . However, some VoIP providers limit the range of free telephony to users who have registered with them or one of their partners. In this case, the user has the option of addressing his call partner directly via the IP address without using a VoIP service provider for call-free telephony.

For calls from the Internet to a subscriber in the classic telephone network, a gateway is required to establish the connection. Its use incurs costs that consist of the provision of the infrastructure and the call charges in the telephone network.

When making international calls to a subscriber in the classic telephone network, the location of the gateway is decisive: the inexpensive Internet access is used up to the gateway, after which the telephone prices of the gateway provider apply.

If an existing company network is used for IP telephony, there are no connection costs that depend on the duration of the call. In addition to the costs for VoIP-capable network components (router and LAN switch ), the proportional costs for the network bandwidth must be included in a profitability analysis. The bandwidth required results from the bandwidth per call, depending on the codec used, and the expected number of simultaneous calls.

Security aspects

The integration of voice data transmission into the IP network poses new challenges for IT security. In its broadcast on February 3, 2015, the ARD magazine Report showed that representatives of the secret services of several countries, including the BND, had already worked with VOIP providers in 2004 to develop “VOIP-LI standards”. "LI" stands for English lawful interception .

The VoIP packets are transmitted via a so-called "shared medium", ie via a network that is shared by several participants and different services . Under certain conditions, attackers may be able to access the data on the transmission path and record the conversation. For example, there are programs with the help of which the data stream can be tapped from switched environments by means of " ARP spoofing " and an audio file can be generated from it.

Although it is possible to encrypt the transmission with Secure Real-Time Transport Protocol ( SRTP ), this is rarely used by users as most VoIP providers do not support it. Another reason is the lack of knowledge about this possibility, and encryption can also impair the voice quality, which is why users often decide in favor of voice quality rather than higher security.

The Session Initiation Protocol (SIP), which is often used, cannot be regarded as sufficiently secure in all forms that are found in practice. Although it does have security mechanisms (for example call IDs based on hash functions ), it offers attack options for denial-of-service attacks.

Another security-relevant area is not exclusively limited to this technology, but is favored by the low costs involved in the calls. There is the possibility of a type of “VoIP spam”, also called SPIT (“Spam over Internet Telephony”).

In vishing , the equivalent of phishing , criminals pretend to call a bank in order to steal the passwords of unsuspecting customers.

In addition, phreaking with VoIP could experience a revival, so to speak. The scenario is based on the fact that in VoIP communication the signaling (for example SIP) is decoupled from the voice data (payload, for example RTP ). Two specially prepared clients set up a call via the SIP proxy and behave in an absolutely standard-compliant manner. After the call has been set up, the SIP proxy is informed that the call has been closed. This sees the session as ended and books the call. However, the RTP data stream is maintained by the clients. The call partners continue to call for free.

VoIP telephone systems (e.g. in company use) and all other VoIP devices that communicate directly via VoIP on the network side require a completely new security assessment. For the sake of simplicity, only the telephone systems are discussed below. The explanations apply in principle to every device that can be reached directly via VoIP on the network side.

While conventional telephone systems could only be reached from the outside via ISDN or analogue lines and only rarely had a connection to the company's internal data network (e.g. for configuration purposes or CTI), VoIP systems that use VoIP on the network side can act as a gateway for new types of hacker attacks.

In order to be reachable for incoming calls, it is essential to open the ports required by VoIP telephony in the firewall and to forward incoming data packets to the telephone system at these ports. Since such packets (= calls) arrive both unsolicited and unplanned, these ports must be permanently open and cannot be triggered by outgoing packets. The system can therefore be reached continuously and unfiltered on these ports.

Modern VoIP systems are often part of the local network - or must be if VoIP devices are also used internally. If a potential attacker were to succeed in taking control of the telephone system by transmitting manipulated VoIP datagrams, for example, he would have gained access to the entire local network. Routers, gateways, servers and similar components are usually checked for such weak points, whereas this aspect hardly required any attention with conventional telephone systems. In the future, VoIP telephone systems will have to be classified in the same way as other devices exposed on the network side under security aspects.


By eliminating the classic telephone lines, the local data network in companies represents a single point of failure for communication between employees. If they could still be reached by telephone without VoIP in the event of a network component such as a switch or router failure, this is no longer the case with VoIP or only to a limited extent via mobile phones. Investing in a redundant network can reduce this risk.

Power supply

In classic (circuit-switching) telephone networks, connections were operated with an exchange remote feed , which supplies the connection with energy independently of the local power supply. While this remote power supply is still sufficient for end devices on analog subscriber lines for full operation, with ISDN for a single end device in emergency mode, it is insufficient for powering devices for operating VoIP (e.g. routers, terminals).

If the VoIP functionality is still to be possible at these connections in the event of a local power failure, then all components, DSL modems, routers, VoIP end devices must be protected by an uninterruptible power supply .

A similar situation exists with many modern analog telephones. Most cordless telephones in particular also do not work without local power supply to the base station.

Localization and emergency calls

Since the telephone number is not necessarily location-specific, the caller can only be localized to a limited extent. This is particularly problematic with emergency calls , where help is very difficult without the appropriate location. It also applies to offers that have geographical dial-in numbers in order to provide region-specific information (directory inquiry services, service or call centers, special numbers).

Public safety and state blockade

Since the telephone numbers are not tied to a specific location, the country code depends solely on the SIP provider. Therefore, the country code (around 49 for Germany) does not tell you where the call is really coming from. According to intelligence sources, terrorists could use VoIP to communicate because of this. So out of is Edward Snowden ge leakten documents seen that the NSA and GCHQ since 2008, various VoIP channels of online games monitored. Statements by the Belgian Interior Minister Jan Jambon were picked up in the media around the Paris terrorist attacks , according to which the IS terrorists are increasingly communicating via party chat , the VoIP feature of the PlayStation 4 . In Arab countries in particular, more and more Internet service providers are blocking IP telephony, such as the Moroccan Maroc Telecom .


  • Kai-Oliver Detken , Evren Eren: VoIP Security - Concepts and solutions for secure VoIP communication. Hanser Verlag, 2007, ISBN 978-3-446-41086-2 .
  • André Liesenfeld: Unified Communication Practical Guide. Plan, implement and use united communication services successfully. Hanser, Munich 2010, ISBN 978-3-446-41834-9 .
  • Andreas Kanbach: SIP - The technology. Vieweg, 2005, ISBN 3-8348-0052-X .
  • Thor Alexander: Internet telephony, VoIP for everyone! Hanser, 2005, ISBN 3-446-40456-2 .
  • Marc Sielemann: Voice over IP. Profitability for large and medium-sized companies. Shaker, 2005, ISBN 3-8322-4591-X .
  • Jochen Nölle: Voice Over IP. Basics, protocols, migration. VDE, 2005, ISBN 3-8007-2850-8 .
  • Anatol Badach: Voice over IP - The technology. 4th, revised edition. Hanser, Munich 2010, ISBN 978-3-446-41772-4 .
  • Egmont Foth: IP telephony, manual. FOSSIL, 2001, ISBN 3-931959-33-3 .
  • Rolf-Dieter Köhler: Voice over IP. mitp, 2001, ISBN 3-8266-4067-5 .
  • Hein, Reisner, Voss: Voice over IP. Make the right use of voice-data convergence. Franzis, Poing 2002, ISBN 3-7723-6686-4 .
  • Jörg Henkel: Voice over IP - Legal and Regulatory Aspects of Internet Telephony . Publishing house Dr. Kovac, Hamburg 2009, ISBN 978-3-8300-4379-9 .


  • RFC 741 Specifications For The Network Voice Protocol (NVP)
  • RFC 3261 SIP: Session Initiation Protocol

Web links

Wikibooks: IP Telephony  - Learning and Teaching Materials
 Wikinews: Category: VoIP  - in the news
Wiktionary: Internet telephony  - explanations of meanings, word origins, synonyms, translations
Wiktionary: Voice over IP  - explanations of meanings, word origins, synonyms, translations
Wiktionary: VoIP  - explanations of meanings, word origins, synonyms, translations
Wiktionary: voipen  - explanations of meanings, word origins , synonyms, translations

Individual evidence

  1. Voice-over-IP - Duden , Bibliographisches Institut ; 2016
  2. [ https://www.bundesnetzagentur.de/SharedDocs/Downloads/DE/Allgemeines/Bundesnetzagentur/Publikationen/Berichte/2017/TB_Telekommunikation20162017.pdf?__blob=publicationFile&v=3 ] (PDF) Bundesnetzagentur, p. 17 (PDF); accessed on January 23, 2018.
  3. ^ Danny Cohen , Stephen Casner, James W. Forgie: A Network Voice Protocol NVP-II . (PDF) ISI / RR-81-90
  4. mtalk and other early Linux VOIP packages
  5. Changeover from 01805 to 032 area code at Deutsche Telekom for value-added service Fax & Fon
  6. Accessibility of 032 numbers .
  7. Vodafone as the last major mobile operator in Germany activates the route to the 032 area code .
  8. 032: Special area code for Internet telephony. Teltarif, accessed on March 15, 2015.
  9. Sample calculation for the required bandwidth
  10. The Good News on VoIP - Reliability Improves on Latest Keynote Study of Internet Telephone Service . Press release, January 25, 2006.
  11. Sagem offers hybrid fax for faxing via Voice over IP. In: golem.de. March 20, 2007, accessed October 30, 2014 .
  12. Daniel Hüfner: Looking for a telephone system? 12 providers in a quick check. Retrieved April 30, 2019 .
  13. ^ Marie Keyworth: Vishing and smishing: The rise of social engineering fraud. In: bbc.com. January 1, 2016, accessed April 10, 2017 .
  14. Jaikumar Vijayan: The NSA tracks World of Warcraft and other online games for terrorist clues. In: computerworld.com. December 9, 2013, accessed May 23, 2016 .
  15. Victoria Ho: There's no evidence ISIS used PlayStation 4 to coordinate the Paris attacks. In: mashable.com. November 16, 2015, accessed May 23, 2016 .
  16. ^ Maroc Telecom Blocks Online Games. In: moroccoworldnews.com. May 20, 2016. Retrieved May 23, 2016 .