CNA Cheat Sheet (DRAFT) by seashore

When the ITSP needs to direct calls to the PSTN, it forwards the calls to a SIP/ PSTN gateway that converts the signaling and voice media to the format the PSTN is expecting

The proxy sends the INVITE on to the SIP/PSTN gateway that will in turn send the approp­riate SETUP message to the PSTN switch and onto the telephone

Once the phone starts ringing, the 'alerting' and '180 ringing' messages are sent back to the SIP device

The media stream starts to flow after the SIP device sent it acknow­led­gements

Both proxy and gateway can be used ambiguous depending on where they're being used and by whom, but ... A proxy generally talks SIP only (i.e. on both sides), and acts as a router for inbound and outbound requests (just like a web proxy). This might be required for security, NAT traversal or other reasons.

Early media is not required on standard PSTN calls as when a number is dialed a media channel is establ­ished so the caller can hear the ringing tone of the remote device.

Clipping is a problem where if a person using a PSTN phone answers their phone and starts talking straight away, without Early Media the SIP phone that is calling them will miss the first part of the conver­station as it hasn't yet received a 200 OK message to enable it to set up the RTP media path.

TRIP is protocol indepe­ndent in that it can be used not only with SIP but also other protocols such as H.323

The gateway tells the Location server of the numbers or routes that they can get to and their own contact details

When a SIP client calls a number in another part of the world, the INVITE goes to the proxy. The Proxy may use ENUM to see if the number has a SIP location. If not, the Proxy forwards the SIP INVITE to the gateway. The gateway will check with the Location server to see which is the best Gateway to breakout to the PSTN. The Location server looks at its table and returns the relevan info to the Gateway. This then contacts the gateway at the remote end and the call is setup via that local gateway

The legacy SS7 ISUP messages have to be interr­ogated by the SIP/PSTN gatway and then the info that will help SIP proxies to route the SIP message is built into the SIP header while other ISUP info is added as a MIME message body. This message can be encrypted.

SIP-I (SIP with encaps­ulated ISUP) was developed by the ITU, not the IETF for SIP-T. It is more accurate than SIP-T

Main transport protocols - TCP and UDP

UDP - user datagram protocol is used to carry voice. There are no acknow­led­gements and no resends. This is the protocol of choice for real time applic­ations such as video and voice.

UDP is becoming obsolete because a lot of SIP products now produce headers that are too big for UDP. TCP is being used more because it has the ability to break up messages, re-ass­emble them at the destin­ation and cope with packet loss with retran­smi­ssions.

Layer 5 Session - SIP messages

TCP segments has a maximum segment size of 1500 bytes.

Usually the network supporting all the components of ah IP PBX is an Ethernet LAN. In some cases, the components can be remotely located over a WAN

Trunk gateway are useful for connecting to the PSTN/POTS network.

A router can be used to connect to other IP PBXs over an IP LAN or even SIP trunks of VoIP service

Signaling path does NOT carry the voice call.

TFTP is used to download software and config­uration inform­ation to the IP phones, softphones and gateways.

This connecton uses the real time protocol (RTP) to carry the voice packets between devices

Encoding converts the orginal analogue signal into a binary data stream.

Digital telephony requires 64000 bits per second for normal speech quality

G.711 - uncomp­ressed voice

It is designed by the IETC codec working group and incorp­orates technology from the Skype 'Silk' codec and Xiph.org CELT codec technology

Implem­enting OPUS in VoIP will allow intero­per­ability with new WebRTC enabled browsers and devices without any transc­oding required

Most instal­lations use packet rate or size of 20ms for the voice or video element when sending a stream of data

Reducing the packet rate means that less work needs to be done on routers when it comes to analyzing IP addresses, checking QoS settings. It does also mans that packet loss can be more apparent to users as there is more inform­ation in a larger packet which if lost, results in a larger gap in a voice stream.

The first two bits of the RTP header donate the version of RTP. ie version 2

The next bit is the extension bit. If it is enabled then RTP header is followed by one extra header extension

The next bit is the marker bit which is used when events such as DTMF tones, volumes and other states that will effect the RTP stream need to be carried in the payload. eg 911 call

The sequence number is used by the receiving device to detect if there is any packet loss and to put received packets in order. The initial value in a session is random and for each RTP packet it is increm­ented by one.

32 bits are used to identify the synchr­oni­zation source. It is a random value with the intention that no two synchr­oni­zation sources within the same RTP session will enven have the same SSRC id.

RTCP carries an ID for a device that is transm­itting the RTP data and this is called a canonical name of CNAME, alias. If the synchr­oni­zation source changes, such as when another stream is beingin introduced from the same device or machine, the CNAME remains the same. This helps to identify partic­ipants in a session and is useful as these partic­ipants can join and leave sessions dynami­cally

RTCP XR is used defined in RFC 3411

The protocol measures VoIP call quality using these following key metrics:

2. Round-trip delay

4. Call quality in terms of estimated R factor or mean opinion score (MOS)

Types of RTCP packet are:

RR- this is the receiver report and it shows the reception statistics from partic­ipants that are not actively sending data in the session

GOODBYE - this indicates end of partic­ipation in a session

XR - an RTCP extension that can help provide a 'rich set' of data for voice management - if it is implem­ented VoIP devices

Delay - one way end to end voice delay should be no more than 150ms

The Intern­ational Teleco­mmu­nic­ation Union Teleco­mmu­nic­ation Standa­rdi­zation Sector (ITU-T) G.114 recomm­end­ation specifies that for good voice quality, no more than 150ms of one-way, end-to-end delay or latency should occur.

Jitter - no more than 30ms

Bandwidth

Network delay happens in various locations such as:

2. Variable delays, eg router queues

4. processing delay in the end devices

Jitter can occur from many of the network problems that cause delay like router queues, over utilized devices, poor cablings

A packet of data leaves a phone at exactly 20 ms intervals using G.729 an on average they arrive at the destin­ation phone every 20 ms. A jitter buffer at the receiving end copes with irregu­lar­ities of the network. This is a buffer that holds each packet just long enough to allow packets to emerge at precisely 30 millis­econd intervals

1. separate traffic for securitiy

3. separate traffic due to different charac­ter­istics, ie data and voice

The switch will tag the frame with a VLAN ID and priority value for its life with the switched network.

A lot of VoIP manufa­cturers set their IP phones voice datagram value to 46, high priority RTP traffic

eg Class 4 (CS4) you see that AF41 means Assured forwar­ding, Class 4 with a drop precedence of 1

Class 4 is higher than class 3. Packets marked AF43 would take priority over packets marked AF31

Bandwidth hogs causing poor VoIP quality

Poor quality wiring network deployed over a number of years. Can cause issues such as delay, packet loss

DNS perfor­mance issues. Critical for VoIP as used for call setup

Router flaps are where constant changes in route path of a call stream results in jitter. This could be caused by traffic engine­ering in place for load balancing.

Port 5060 is the default UDP port for SIP commun­ication

SDP wireshark eg audio 20016 - media is audio and the UDP port is 20016 that the device is advert­ising to receive Voice or RTP packets. Another media type is video

SDP wireshark eg rtpmap:8 PCMA/8000- The rtpmap element describes a little more detail of the codecs being advert­ised. 8 is the G.711 Alaw codec samples at 8000Hz

The offical format of rtpmap is [a=rtpmap: <pa­yload type> <en­coding name>/­<clock rate>{­/<e­ncoding parame­ter­s>

eg audio 56828 - the remote SIP device advertises the UDP port that it want to receive RTP audio on.

8 is G.711 ALaw

SDP doesn't have to understand fmtp:18 annexb=no, it simply passes the inform­ation between SIP devices and hopes that they unders­tand.

SIP devices do not have to send RTP packets if there is nothing to send. eg silence in a conver­sation

The nortprpoxy attribute tells us if the remote SIP device is using an RTP Proxy.

RTSP runs over UDP, therefore there is not mechanism to compensate for packet loss

High packet loss over 5% can cause noticeable picture distor­tion. Lost packets can also affect the sound quality.

Smaller jitter buffers will force IP network designers to keep jitter in check on the network.

A larger jitter buffer could solve the problem but add more end-to-end delay which is not accept­able.

Video codecs - VP8, VP9 free

Resolution - standard, HD=192­0x1080, ultra HD, 4K

QoS tagging video is important

m:audio for voice, eg codecs Opus, SILK

processing authen­tic­ation requests

Unders­tanding / mainta­ining other services being used by the UA which might need to be taken into consid­eration when applying AS-SIP capabi­lities

Working with the access router - establish permission at the access router for it to handle the precedence marked packets from the UA

Maintain all records of the service, whether for accoun­ting, auditing, or other purposes.

Transfer packets - decides which packets are to be transp­orted between networks or domains. If access is not granted, the access router will throw these packets away

Resour­ce-­pri­ority header contains namespace and priority known as the r-priority values.

200 OK message will have Accept­-Re­sou­rce­-Pr­iority along with the namespace and r-priority value

eg Accept­-Re­sou­rce­-Pr­iority: q735.0, q.735.1, q.735.2, q.735.3, q.735.4.

-Proxy receives message

-Does recognizes and authorized it uses the priority levels. If it does not know the namespace and the UA has been authorized it will use the priority levels in the message

-If UA doesn't have a priority level set, the proxy can assign a default

-If auth and resource not available- priority is utilized.

MLPP is built as proactive system in which callers must assign a precedence level at call initia­tion, this precedence level cannot be changed throughout that call.

When any trunk config­uration reaches its capacity, a choice must be made as to which call can continue.

The main elements of MLPP are: A VoIP implem­ent­ation of an MLPP service must provide these charac­ter­istics.

All callers that have been pre-empted to make way for the new call must be informed that their call has been pre-em­pted, and have to make way for the higher precedence call.

Author­ization of call placed Calls can be made using a higher level of priority by either inputting a specific priority code before a number or setting variable on the SIP UA via policies so that the caller always has a minimum priority level