I'm not able to find clear explanations of how VOIP works

Posts: 10 · Vonage Forum Associate Joined: May 21, 2005 Posts: 10

Can someone please tell me how Voip works. I've got contradictory explanations even from Vonage people. One version is it is a steady signal with its own frequency, and one version is it is IP packets indistinguashable from e-mail except for the volume, and yet another is it specifically needs unique ports because it has its own frequency. Yet another version is that the reason why it won't interfere with internet use in the house is it has its own frequency; yet I read alot of user reports of internet phone and other big internet operations interfering with each other because they both use packets and bandwidth.

What does the signal consist of, how is it really transmitted, and why does it need to be transmitted through special ports?

All I can find on the internet is vague statements about IP and packets and nice kindergarten drawings of lines from different elemnts of the system. That includes cnet articles, Voip 101, everything. Explanation on the Vonage web site is the least explanatory of all!

Can someone please provide or point me to a clearer explanation. Thanks!

Strange - you'd think such a growing technology would be explained more simply than that! I didn't have trouble learning what an IP address and a url are, nor how e-mail travels, nor what happens when nationwide e-mail servers and lines back up!

Yours,
Dora

Posts: 2251 · Posted: Sat May 21, 2005 2:56 pm Post subject:

This file was sent to me in a email about three years ago. I do not recall who sent it or who the author of the document is. It might help you understand SIP, which is the Voip protocol that Vonage uses.

How does SIP work?
An indepth look at SIP

Standard telephony has been with us for a long time and is the most reliable form of communication. The PSTN is a network built specially for this communication method, and so contains all of the signalling and call set-up protocols.

Voip is poised to start replacing this network by running everything over the data network. The potential cost savings make this a very attractive proposal, but there are still problems - the data network was never designed to handle complex communications sessions.

In Voip, none of the signalling protocols required are present,

and a user may be constantly switching IP address, either by physically transferring his or her location or because of the expiry of DHCP leases.

This immediately makes it more complex to make connections. So, in this week's RTFM we'll take a look at the Session Initialisation Protocol (SIP) and see how it can help to address these problems.

What does it do?

SIP is an IETF standard designed to provide advanced telephony services over the internet. As the name suggests, it is used to establish, terminate and modify communications sessions in an IP network. It began life as a component of the multicast backbone (Mbone) network. This was an experimental network designed to deliver multimedia content over the network.

SIP is not a standalone protocol and doesn't actually understand what a session is. Instead it interoperates with existing protocols and just swaps Meta data that enables session initialisation.

A session can be as simple as a two-way telephone call, but can also be a fully-blown multimedia conference call.

As SIP only provides the method for dealing with sessions, it has to work alongside other protocols and standards, which provide the level of service required by real-time communication. Typically, communications require a guaranteed delivery time, which can be provided by the real-time transport protocol (RTP). Voice quality has to be guaranteed using protocols such as RSVP and YESSIR. Directory integration with LDAP is essential for user discovery, while authentication server support, such as RADIUS, ensures that users are correctly identified.

What does it provide?

To facilitate proper communication, SIP has to provide a range of features and functions to establish a session. We'll take a look at these in order and talk about the associated problems.

l User Location

SIP has to provide the function that locates where a user is currently located. It is common for a user to have a PC at work and home, a laptop and even a Voip phone. An incoming connection might require that all devices ring at the same time, that one device gets preference or that a round-robin ring around takes place. SIP has to deal with this dynamic configuration information and correctly locate the user.

Call Initialisation
Once a user has been located, a session has to be established. Although SIP doesn't understand the session, it has to transmit the description of the session from the caller to the receiver. This usually results in negotiation between all of the parties involved in the call. It could be that a device doesn't support video, but only voice.

SIP transfers this information using the multipurpose internet mail extensions (MIME). This is used in web and email services to describe content. This means there is always a wide-range of options available in negotiation. The most common form of session description used by SIP is the Session Description Protocol (SDP).

Call modification
Once a call is established, SIP still has a role to play. A call in progress can change features, such as a video stream being added to a voice conversation. New parties can also be added into a conversation, which requires additional negotiation of features.

Call modification doesn't have to be as drastic as either of these measures. Common telephony functions, including muting and call holding, are also supported by SIP. The range of call features supported depends on the session description protocol in use.

Call TerminationAfter all users are finished with the session they hang-up. SIP has to deal with this call termination.
How does it work?

Now we understand the facilities that SIP provides, we can look at how it provides them using a client/server architecture.

User Agent Client

The User Agent Client (UAC) sits on the client and is responsible for making any SIP requests to other clients. The UAC is also the component that accepts and deals with any request that comes in, including call initialisation, modification and termination.

User Agent Server

There are three types of user agent server (UAS): stateful proxy server, stateless proxy server and the re-direct server.

Proxy servers receive request and work out where to send them. This is generally onto the next server until the destination is located.

A stateful proxy server retains information of all requests and responses. This can be used to send a request to multiple destinations and only pass on the best response. A stateless proxy server simply passes requests on but doesn't retain any information.

A re-direct server doesn't proxy requests, but returns the destination address to the caller. The caller can then make a request directly to the destination.

How does this work?

We can see how this works by taking a look at a simple example of user 'A' trying to call user 'B'. A's UAC sends a SIP invite request to its local UAS. If this is a proxy server, then it forwards the request to the next UAS until it hits its destination. If it is a re-direct server then it returns the direct address of B. In SIP, addresses are represented as URLs and follow a similar layout to email addresses.

The invite includes a full description of the session including all of the media streams that A wants to use. B replies to the invitation, but includes a description of any modifications that he wants to make. This is for compatibility reasons, as B might support all of the features that A asked for.

After this negotiation is completed, the session is created and A and B can communicate. At the end of the call, either side can send a disconnect, terminating the session.

All of this process is automatic. For example, when A calls B, if B picks up his Voip videophone, the phone automatically handles the media negotiation process. When B puts the phone down, the disconnect is automatically sent.

It is also the job of a SIP device to register its current location with a UAS. This ensures that a user can be found even when mobile.

Advanced Features

Earlier in this piece we discussed the facilities that SIP provides. One of these was user location by simultaneously ringing all of a user's devices. In the simple example above we only cover a single connection. However, SIP can use a process called 'forking', which sends out invites to multiple devices at once. The first device to respond gets the connection. This is similar in conception to a phone pool in a helpdesk, where all incoming calls ring all available phones. However, this feature is only available if a stateful proxy server is used, as it needs to remember which connections to allow and which to block.

SIP can return different media types. If a client connects to a company, then they could return a list of phone numbers for the building. It means that information can be dynamically created and transferred.

SIP Messages

SIP has well-defined messages that are used for communication. A message can either be a request or a response message. We'll take a look at each message type and the options that are available.

Request Methods

INVITE is used to initiate a call and is also used to change call parameters once a session has been established. This can be thought of as a re-INVITE.
ACK is an acknowledgement and confirms response to an invite.
BYE terminates a call in progress.
CANCEL Terminates the invite and stops all searches and devices that are ringing.
OPTIONS Used to query the capabilities of a device.
REGISTER Registers a device's location with a server.
INFO Sends information during a session but does not modify the session state.
Response Types

Response messages contain numerical responses and are similar to those used by HTTP:

1xx provisional, searching, ringing, queuing
2xx success
3xx redirection, forwarding
4xx request failure (client)
5xx server failures
6xx global failure (busy, refusal, not available anywhere
SIP Message architecture
Both request and response SIP messages are built from the same three components:

Start Lines
The Start Line is the beginning of every SIP message. It is first used to identify the type of message and the protocol. Depending on the message type the Start Line is either a Request-line for requests or a Status-line for responses.

A Request-line includes a Request URL, which is used to identify the requested user. This line can be rewritten by proxies, although the 'To' field later on cannot.

A Status-line holds the numeric Status-code and the text version.

Headers
SIP headers contain a list of fields that are used to identify message attributes. The syntax is:&ltfield>:&ltvalue>

Body
The body conveys the mechanics of a connection. It is common for the body to hold protocol information such as SDP, and only the receiving client can understand what the body means.

Conclusion
SIP provides an intelligent method for introducing advanced call features to an IP network. Its protocol independence means that it can be used to initialising calls with features ranging from instant messaging to voice calls.

Just remember is that SIP works in the background and makes IP communications just as easy as picking up the phone.

Response codes

Unavailable100 Continue
180 Ringing
200 OK
301 Moved permanently
302 Moved temporarily
400 Bad request
408 Request time-out
480
600 Busy
603 Decline
604 Does not exist

Call and effect: SIP fields

Invite Used to address an invite
To Address of user being invited
From Address of user originating request
Subject Call subject

Posts: 646 · Posted: Sat May 21, 2005 4:03 pm Post subject:

That SIP stuff describes how calls get set up, but not so much about the fundamentals of where your voice actually goes.

The stuff you heard about "different frequencies" is probably in relation to some of the digital phone services offered by cable companies. Those services aren't really Voip, because they just use the cable TV lines to send your voice locally, and don't really have anything to do with the Internet.

Vonage, on the other hand, is real Voip. You plug your phone device into any existing Internet connection, and the voice gets chopped up into 20 millisecond blocks, and sent as individual packets over the Internet.

When you call a land-line phone, your voice packets travel to one of Vonage's gateway servers (I think mine's in Chicago), and then hook into the normal phone system from there. So, when I call a phone next door, the voice has to go a few hundred miles, and that causes some lag.

I think that when you do a Vonage-to-Vonage call, the voice packets are sent directly from one user to the other, without getting routed through any Vonage servers. So, for example, if you call Bob next door, and you both have cable modems, then the voice packets stay within the cable network and never have to go more than a few miles. [can anyone confirm this?]

Oh, and that stuff about ports... that's referring to TCP and UDP ports, which are numbers that computers on the Internet use to keep track of connections. It's not actual plugs in the hardware.

Posts: 134 · Vonage Forum Senior Joined: Mar 23, 2005 Posts: 134

If you saw my recent post it gives you an idea of how it works as far as Vonage is concerned.

When you call your neighbor your Vonage device initiates a SIP message to Vonage's servers containing information about your call such as the number, your device MAC, IP, etc. Because of my location my calls are routed to Vonage in New Jersey via the internet. This server only handles the SIP traffic and basically controls the call. This server along with your Vonage device are in constant communication with each other on port 1000, even if you are not calling someone. This server then decides what route and carrier are to be used for your call. Vonage contracts with two or three different carriers who have private networks setup across the country and world for that matter. When the server makes the decison of where to route your call, that carrier's server responds and connects to two ports on your device. One is for upstream and the other for downstream. Your call then traverses the carrier's private network to a PSTN (Verizion, SBC, etc.) gateway located near the person you are calling. Your call then enters the PSTN network and you are now connected to your party. When you hang up, your device sends a terminate SIP message to the Vonage server overseeing your call. The server then tells the carrier' s server to terminate the connection.

I don't understand how Vonage to Vonage calls work. But I imagine it's a similar set-up to what paul said. My confusion comes from not completely understanding the variables involved with Vonage calls such as calling a Vonage number that was previously associated with a landline phone.

Posts: 1 · New Forum Member Joined: May 21, 2005 Posts: 1

Simple way to explain it:

When you visit a website, you open a request to the server via port 80. The packets contain commands, such as "give me the default page", etc.

Voip sends packets in the SAME exact manner as a webpage request does, only it contains your voice. The Voip machine that you hook into your internet connection will send/receive this data packets. It sends the data packets to Vonage, and they do whatever they do to get it onto the normal phone system.

So you see, the data is a lot like e-mail, web browsing, downloading, etc. It's just... voice. Very Happy

And like other things, if you're downloading a lot of files, your internet goes down, there is a problem at the cable company, etc, your phone can go out.

Please note that there is no such thing as 'frequency' on internet connections. (There actually is, but it's on a much lower level, nothing the end user needs to worry about) It's all data, just on diffrent ports. (Like an shipyard with diffrent peers, or 'ports' as you would have it!)

Hope that explains it!

Posts: 90 · Vonage Forum Senior Joined: May 05, 2005 Posts: 90

It's all done by little elves....

Who cares, so long as it works?

(it is basically like any other data you up/download from the 'net, be it text, pictures, video, audio, etc. The only difference is that the Vonage box emulates an analog phone system on one end).