Category: Wireless VoIP

This is not a very mobile blog entry today unless you are using Skype via one of the third party clients now available. For many years I’ve heard people in the industry talking about how to integrate telephony functionality into web pages. Never saw a web site that actually did it until today. Take a look at the picture on the left. It’s part of a web page of Hotel Morgana in Rome, Italy. A Skype call button and the invitation to call them via Skype for free is placed neatly in the middle of the page. It makes total sense for them as they have guests from all over the world who would not be able to call national 0800 numbers. For them it also makes sense as the cost of offering the service for them is either zero as well or only very small in case they use Skype call forwarding to their fixed line phone. This is free for the caller and only costs a couple of cents to the hotel owner.

One of my favorite topics is cellular network capacity. I posted an example a while back on the 1 kb/s 3G surfer. At the time I excluded VoIP as part of my application mix as it increases network traffic quite a bit. Today I found some interesting material which sheds some light on this part of the story:

Wireless Networks have a particular problem with Voice over IP. While traditional circuit switched traffic was optimized on all layers of the protocol stack to be transferred as efficiently over the air interface as possible, achieving the same effect for Voice over IP is very difficult due to the decoupling of the different network layers on the IP protocol stack. Thus, a VoIP call today consumes at least four times as much bandwidth on the air interface than a circuit switched voice call. In other words, if everybody started to use VoIP over wireless today, network capacity for voice calls would shrink to only a quarter of what it is today.

Now Ericsson has released an interesting slidepack on the topic which targets the non geek investment community but which nevertheless contains some interesting numbers on wireless network enhancements and optimization for VoIP in the future: As described above, Ericsson’s slide pack shows on page three that  UMTS networks today could only provide 20% of the voice capacity with VoIP compared to standard voice calls. So that’s close to my number above. With HSPA (I think they  refer to HSDPA + HSUPA) capacity increases to 70% if IP robust header compression (ROHC) is used. They then go on to claim that with efficient signaling (whatever that is…) and improved scheduler (what improvements?) and optional GRAKE2 (again what is that?) VoIP capacity can be pushed to 140% of today’s standard voice capacity. Finally the next evolution of UMTS called LTE (Long Term Evolution) targets 200% of current voice traffic for VoIP in the same bandwidth.

Not sure what the improvements are they are talking about as they don’t give any further explanations but it seems we are getting somewhere in due time. VoIP VoIP Hurray!

This blog entry is part four in my mini series of looking at the different Voice over IP systems (VoIP) that can be used over wireless networks such as UMTS or CDMA 1xEV-DO. Part 1 focused on UMA, part 2 on SIP, part 3 on IMS, and this part will take a look on the use of Skype over wireless.

UMA, SIP and IMS are all centralized systems. That means that they use a centralized server which is responsible for authenticating users, for establishing connections between users for voice, video, instant messaging or any other kind of media transfer, and also for billing. Skype uses a fundamentally different architecture as it does not rely on a centralized server for most of these tasks. Skype is a Peer to Peer network in which end points of the network help out each other to establish and maintain a connection.

A peer to peer network has a number of advantages over a centralized approach:

• Centralized servers are costly to buy, maintain and operate. The more people use the service the bigger the server has to become. In a peer to peer network such as Skype, however, signaling load at a central point does not increase in the same way as in centralized systems.
• Individual peers help out each other to establish a connection. This is especially important as many users are behind firewalls or network address translation (NAT) routers typically used at home. Thus, they can not communicate directly with each other. Skype peers that have no such restrictions help out peers that do and forward traffic between such users. This is the main reason why Skype is so easy to set up on PCs and other devices compared to other technologies like for example SIP. For those of you who would like to find out more about Skype, here’s a link to an analysis of how Skype works which has been published by Philippe Biondi and Fabrice Desclaux of EADS.

While most other VoIP systems use legacy voice codecs to transport the media stream over IP, Skype uses its own resource efficient codecs which which on top even have a superior voice quality. Thus, Skype works quite well over UMTS and I use it on a regular basis when traveling. It should also work quite well over EV-DO as well, as bandwidth is also sufficient. Personally I’ve never tried so this is just a speculation.

Many operators (carriers) are scared of Skype and other VoIP systems as they are afraid that such services will decrease their revenue on traditional voice minutes. I think there is no such risk in the near future as there is still a PC required to run Skype over a wireless link. However there are first signs that Skype is also moving to mobile devices. A beta client for Windows Mobile is already available and a non official beta of Skype on a Nokia S60 6680 has also been spotted by the author (see picture above). So operators should hurry up and develop strategies to integrate such innovative applications into their concepts. Some have already done so, like for example E-Plus in Germany. They even offer a UMTS flatrate together with the Skype software and a headset. An interesting first step, certainly not made too soon as new devices such as the Nokia N80 with built in WLAN will spur the interest of a wider audience to cheap VoIP over wireless.

At this point I close my wireless VoIP mini series for now. Four different VoIP systems, four different basic ideas and four ways for every one in the industry and of course the users to benefit. I think it will still take several years before Wireless VoIP becomes mainstream but the first signs are already here.

Wireless VoIP is a hot topic at the moment and the name is used for quite a number of different technologies. In this mini series on my blog I’ve already taken a look at UMA in part one and SIP in part two. This entry takes a closer look at the IMS, the IP Multimedia Subsystem.

In very generic terms, SIP as described in part two is the core of the IMS. In addition, the IMS standardizes a number of additional functionalities:

• SIP is the abbreviation for Session Initiation Protocol. As it is just a session (e.g. voice call) initiation protocol, it does not contain any mechanisms to ensure Quality of Service or a certain bandwidth of the connection. While fixed networks usually have enough bandwidth available to ensure the quality of a call, things are  different in mobile networks. Here, bandwidth to and from a subscriber is limited. In addition, the total bandwidth of a base station that has to be shared by all users of the base station is also limited and much scarcer than in the fixed line world. To ensure that a voice or video call established with an IMS capable device is maintained with a good quality of service (e.g. jitter, latency, etc.) and that a sufficient amount of bandwidth is ensured for the call while it is established, the IMS contains mechanisms to communicate with the radio network to ensure enough bandwidth is reserved during the call.
• IMS standardizes authentication and encryption of SIP commands and responses in the network.
• IMS standardizes how media gateways are controlled to enable SIP clients to establish connections with legacy circuit switched fixed line and wireless clients.
• IMS standardizes the access to the Home Subscriber Server (HSS, aka HLR) in the GSM/UMTS/CDMA network.
• IMS standardizes what kind of information is put on the SIM card of the subscriber and in which way.
• IMS standardizes the codecs used for different services (e.g. voice codecs, video codecs)
• And very important: IMS standardizes a service framework and a protocol between the IMS core and application servers that allows third parties to create new services based on IMS and SIP commands. Such services are for example instant messaging, presence, voice mail, video mail, location based services, short dialing numbers, private dialing plans, music and video streaming, push to talk, etc.
• And last but not least: IMS standardizes how to get billing information because after all, operators want to make money with it.

By standardizing all of these aspects it is ensured that an operator can buy different parts of the overall IMS solution from different vendors. In turn, the architecture also ensures that application developers can design new applications for a standard platform and can then sell their products without customization to many different network operators. Furthermore, this approach ensures that IMS systems of different operators can interoperate with each other. This is a fundamental requirement to enable calls between different national and international operators without using media gateways.

Want to know more? Here’s a good book on the topic that I can recommend as it was great fun reading it: "The 3G IP Multimedia Subsystem" by Gonzalo Camarillo and Miguel A. Garcia-Martin. It’s in it’s second edition already after only a year so I guess I am not the only one who likes it.

In the first part of this mini-series on Mobile VoIP, I’ve taken a look at UMA as one of the many flavours of Mobile Voice over IP (VoIP). This part is dedicated to the Session Initiation Protocol (SIP) architecture and how it can be used in mobile devices.

SIP Architecture

SIP is a true end to end VoIP system and has been initially conceived without mobile networks in mind. Today, it is already widely adopted in the ‘classic’ Internet as one of the main protocols for VoIP. The basic architecture of SIP is simple: A SIP server in the network forms the central element of a VoIP network. VoIP clients are either software clients on PCs or notebooks or standalone devices like phones with an Ethernet port. When connected to an IP network their first task is to register their IP address with the SIP server. To call another device, a SIP client then sends a request to the SIP server. The SIP server then retrieves the IP address of the destination device from its database and contacts the destination device. If the destination device is willing to accept the call, the server informs the originator of the IP address of the destination device and a direct IP connection is established between the two parties. For VoIP, the connection is used to carry a voice data stream. Note, that the SIP server is only required for signalling as the voice stream is directly exchanged between the two parties. This is a big difference to circuit switched voice calls which are always routed through a switching center of a fixed or mobile operator. To call non VoIP destinations, gateways are used to convert the IP voice data stream into a 64 kbit/s circuit switched voice data stream.

SIP is more than just Voice Telephony

As the name Session Initiation Protocol (SIP) already implies, it is not limited to setting up voice sessions. More sophisticated clients also use the protocol for video telephony or the exchange of other multimedia data such as Instant Messaging, pictures and file transfers in general. These functionalities, however, are only available if both clients support them.

Comparison to UMA

While UMA (see part one of this series) only replaces the GSM air interface with Wireless LAN and IP, SIP is a true end to end voice over IP system offering a wide range of applications from voice, to video, and to rich multimedia applications like instant messaging and file exchange. Also unlike in UMA, the user is not bound to the mobile operator but can select from an abundant number of SIP server operators on the Internet.

SIP in the Wireless World

In the wireless world, SIP has not been very popular so far due to a number of wireless network limitations. GPRS and other first generation wireless IP packet networks are too slow and the latency of the connection was too high. In addition, speech algorithms used by current SIP implementations use inefficient codecs which require a substantial amount of bandwidth. 3G networks such as UMTS offer higher bandwidths compared to earlier networks and are thus able to carry SIP voice calls over the air interface. A SIP call, however, uses around five times more bandwidth then a traditional circuit switched mobile voice call for which very bandwidth efficient codecs are used in the radio network. This fact together with the openness of SIP for the user to choose the operator of the SIP server himself explains the reluctance of wireless operators to support the application of SIP services in their 3G networks.

The Future of SIP in the Wireless World

In the near future, SIP clients will mostly be adopted on GSM/UMTS/Wifi smart phones such as the Nokia N80, where they can be used to make phone calls over a Wireless LAN access point connected to DSL or a company network. In effect, a SIP client in the mobile phone can replace the fixed line phone at home and I am actually waiting for the day when I can use a single phone at home and when underway.

When leaving the office or home, a SIP client can still be used for voice calls but many operators (carriers) try to restrict SIP for the reasons discussed above. Very near term evolutions of 3G networks to technologies like HSDPA (High Speed Data Packet Access) might change these policies in the mid term.

SIP also forms the core of the IMS (IP Multimedia Subsystem), a standard designed by the mobile industry for mobile operators to offer their own VoIP and multimedia applications. IMS will be the focus of part 3 of this mini-series, so stay tuned…

Nokia has managed quite successfully to bring UMA (Unlicensed Mobile Access), it’s preferred Wireless Voice over IP variant to the attention of the big press during the recent 3GSM congress in Barcelona. Most journalists, however, haven’t really understood what UMA is about as it is just one of at least four very different flavors of Wireless VoIP, each with its advantages, disadvantages, usage scenarios, and proponents that push the solution. To fill this gap and to show the benefits and drawbacks of the different flavors of Wireless VoIP for the user, I’ve decided to write a couple of blog entries in the days to come to compare the following technologies:

• UMA (Unlicensed Mobile Access)
• SIP (Session Initiation Protocol) implementations on mobile phones
• IMS (IP Multimedia Subsystem) clients
• Non standard Wireless VoIP systems such as Skype

First on the list is UMA, a 3GPP standard like GSM and UMTS, loved and feared alike by mobile operators (or carriers as you say in the U.S.). The principle of UMA is simple: It replaces the GSM radio technology on the lower protocol layers of the mobile phone with Wireless LAN. A call is then tunneled via a Wifi Access Point connected to a DSL/cable modem via the Internet and a gateway to the Mobile Switching Center (MSC) of a mobile network operator. The gateway between the Internet and the network of the mobile operator is called a UMA Network Controller and one of the companies developing such a network node is Kineto Wireless.

For me, UMA is a semi-VoIP service, as a call is only transported over IP on the link between the mobile phone and the UMA Network Controller. After the gateway, a traditional Mobile Switching Center (MSC) and a circuit switched connection is used to connect the call to the destination.

By always traversing the core network and an MSC of a mobile operator, UMA binds a mobile subscriber to his mobile network operator. This is the part mobile operators like most about UMA. What mobile operators don’t like about UMA in many cases is the fact that the DSL or cable connection is usually in the hands of other companies. In many cases users pay their DSL fees to an incumbent fixed line operator or cable company. Thus, in most cases UMA only makes sense if a mobile operator offers the service together with the fixed line operator that controls the DSL or cable access.

As described above, UMA replaces one radio technology with another and otherwise leaves the rest of the system as it is. This makes it difficult to price incoming calls differently for a caller while the called party is at home and using his (cheaper) Wifi/DSL/cable connection compared to calls the called party receives while roaming in the cellular network. This is due to the fact that mobile operators in Europe use special national destination codes in order to be able to charge a caller a different tariff for calls to a mobile phone user. In the U.S. charging incoming calls to a UMA user differently might be less of an issue as mobile networks use the same national destination codes as fixed line operators. There is no additional charge for the caller as the mobile phone user gets charged for incoming calls. As the mobile network is aware that the user is currently in his (cheaper) home Wifi cell, incoming calls can be charged accordingly.

Outgoing calls made via the Wifi access point and a DSL or cable connection are also under the control of the mobile operator. It is unlikely that mobile operators will offer outgoing calls for free as is usually the case for connections between two VoIP subscribers as the call will always be routed through a mobile switching center and a circuit switched connection instead of being transported via IP end to end. Consequently I think it’s going to be difficult for an operator to price the service competitively.

Last point on the downside for the user: As UMA is not an end to end VoIP technology there is no presence information and built in instant messaging capabilities as in other systems.

On the positive side, UMA offers a seamless experience for the user. From an application point of view UMA it is transparent to the user on the mobile as the same graphical user interface is used for both cellular and Wifi calls. The standard even offers seamless roaming between the two access technologies for ongoing calls, i.e. a call is handed over from Wifi to the cellular network when a user leaves the coverage area of a Wifi access point.

UMA also tunnels GPRS services into the core network of the mobile operator. Data speeds are much higher though, again producing a seamless or even better experience for the user while in a UMA Wifi cell, e.g. for web browsing on the phone, operator portal access, music downloads, etc.

Two other important positive sides of the technology are the use of the same phone number regardless of whether the phone is connected via the cellular network or Wifi and the fact that indoor coverage can be improved by deploying Wifi access points instead of more expensive cellular micro base stations.

As has been shown, UMA offers a lot to users. If operators figure out a way to offer the service at a competitive price users will surely like it.

In the next blog entry in this series I will give an overview of mobile SIP and how this end-to-end VoIP technology compares to UMA.