220 likes | 303 Views
Global Sales Meeting The Rise of the Mobile Internet Rev 2. Mobile Internet Summit Requirements (for discussion) 19 November 2008. Mobile Internet Summit What are we trying to do here….
E N D
Global Sales MeetingThe Rise of the Mobile InternetRev 2 Mobile Internet Summit Requirements (for discussion) 19 November 2008
Mobile Internet SummitWhat are we trying to do here…. • We can look at how the Internet will have to evolve to address the challenges of a mobile world. The evolution of the Internet is an area where we are expected to lead. • If Vodafone asked us to tell them what the Internet will look like in 3 or 4 years, what would we tell them? • If TMO asked us how they’d make money as a mobile ISP what would we tell them? • What do the BU/TGs need to be thinking about AND building? What do WE want to see happen here? • We could look at problems associated with scaling today’s radio access networks. • This is an important subject as today’s networks are being overwhelmed with data traffic and the traditional RAN vendors are ill suited to addressing these sorts of problems.
Mobile Internet Requirements • This next section will look at some of the requirements that mobility will place on the Internet • It is by no means complete, and we are looking for more input here. • If we can really flesh this out in the next several weeks that would be a big step forward
1) Billions of new users The good news is that this problem is bounded by the number of people on the planet • We will see billions of new users coming onto the Internet in the next decade and most will be mobile attached. • For much of the world the Internet will be a mobile experience 4B Subs 58% 3.5B Subs Global Penetration 51% 2.2B Subs 34% 2011 End 2005 End 2008 End
2) Hundreds of billions of new “things” • The “Internet of Things” is also on the way • We could see tens of billions (maybe even hundreds of billions) of “things” that need to connect • Most will connect over mobile (or wireless) networks • Does the present any special challenges for Cisco and the industry? • NOTE: Today’s networks aren’t designed for this eventuality and it does put an additional load on the GGSN/SGSNs
700 600 500 400 300 200 100 0 3) Radio access options • Huge growth in mobile data traffic will be a challenge for the RF domain • RF is “the major bottleneck” as it just doesn’t scale very well • We will need to leverage ALL available RF tech & spectrum • Licensed and unlicensed • We must make Wi-Fi as easy to use as Cellular • We also need to leverage local area licensed RF technologies like femtocells Worldwide Mobile Data In PB per Month Source: In-Stat August 2008
Wi-Fi and the ease of use challenge • If we want Wi-Fi to be a bigger part of the Mobile Internet experience we must fix the rather horrid Wi-Fi connection management experience • There is just too much of a burden placed on the users • The IEEE 802.11u committee is working on part of the problem by allowing the beacon to advertise additional services • Would allow the beacon to advertise roaming arrangements with aggregators like iPass • A sufficiently clever device can use this to automatically connect a user • Could also allow context to be transferred using a protocol such as the one developed by the IETFSEAMOBYWG. • I expect there are other issues as well • Power control comes to mind
4) Extremely cost effective deployments Traffic & Revenue Challenge • We need to make networks a lot less expensive to build and operate on a cost per bit basis • Operators are getting paid roughly $50 a month to carry around 40 Mbytes worth of circuit switched voice traffic. In the data world we will be looking at $10 per Gigabyte per month in the very near future. That is 25 times the traffic for 1/5th the money • If it’s a gigabyte a month this year it may well be 5 gigabytes a month in 3 years Mobile operatorrevenue & trafficde-coupled Traffic Voice Dominant Revenues Data Dominant Time Source: Unstrung Insider, Mobile Backhaul and Cell Site Aggregation, Feb 2007
5) Simplicity • Today’s radio access networks are way to complicated. Much of this is a function of that fact that data networks are a bolt-on to the circuit switched voice network. • As we go all-IP e2e we need to streamline the network • Let’s boil it all down to access points, switches and routers, and servers • Streamlining of network architecture addresses a number of issues including: • Lower cost from fewer boxes to buy and maintain • Lower latency • More robust networks with fewer boxes that can break
6) Open everything • We must resist the desire of the RAN vendors to build access networks that are closed and proprietary • We need to push for Industry plugfests • Breakup the vertically integrated supply model • Get behind open-source mobile platforms (Symbian and Android) • Insist that ALL interfaces are open and interoperable • Etc.
7) Optimized routing • We need to optimize routing where appropriate • NOTE: In today’s networks everything is hauled back to a tunnel termination point back in a data center. This could be hundreds of miles away from the user. This introduces higher cost, greater latency, and more opportunities for outages. • We should look at methods to optimize that routing where appropriate • Optimization of routing may impact mobility • It is time to rethink mobility • Are there ways to support mobility that don’t use tunnels…. are there ways to use tunnels and still offer optimize routing?
8) Rethinking mobility • Operators are starting to rethink mobility • There is an overwhelming movement to mobile (and wireless) access in the user base, but that doesn’t necessarily mean that the user is actually moving • Building networks for vehicular mobility is an expensive and often unnecessary solution for many situations. • NOTE: >80% of mobile calls are placed from indoors (40% from the home) • Select the right radio access solution for the situation (user, location, and application • Location tracking can play a big role here
9) Mobile devices • High-end devices will support a variety of licensed and unlicensed RF technologies • A device needs to know about RF options without powering up other radios (discovery protocol of some sort) • Seamless movement of sessions across different RF technologies would be a major plus… under control of the user or the network • What is Cisco’s role on the device? CCX, VPN, UC, Connection Mgmt, etc. • This is important as these new mobile platforms will be how people consume the Mobile Internet
10) Multiple device challenge • Allow users to seamlessly transfer sessions between different devices. Goes to the very cool VTG swiping demo on the iPhone. • This will undoubtedly become a compelling application moving forward, but is most likely NOT a network problem. • Let’s assume that the application will take care of this by giving the appearance of seamless mobility by bringing back their connection to the same state on the new device
11) Optimizing the radio link • Mobile operators like applications to be as efficient as possible with their use of radio resources. Packets are not as big an issue as utilization of airlink • It is becoming more of an issue as all sorts of companies develop apps for the iPhone, Android, Symbian, RIM, and WinMo platforms • What can we do to improve airlink efficiency? Reduce the use of chatty protocols, etc. • This is one of the things operators love about RIM. It is very highly optimized for airlink efficiency
12) Service control • Service control is always a big issue with mobile operators. They have been pursuing IMS (IP Multimedia Subsystem) as a solution for years, but this has issues with complexity and lack of a compelling user experience. • Does Cisco want to offer up a much simpler solution here? An IMS-Lite of some sort.
13) Addressing • What about IPv6 vs. IPv4? • Will they both be broadly deployed? Will new apps start to make NAT a nonstarter? • AT&T is wants to make the move to IPv6 starting in 2010. • Other mobile operators are also starting to talk aggressively about IPv6 as well. • What do we need to be doing here to stay in front of this trend? • Are there serious issues associated with actually deploying IPv6? • Do we need to start on IPv9? • Is IPv6 really in their best interest?
What interesting things can we do by exposing mobility to higher layers? Easiest way to do this is to do away with tunnels Higher layer protocols would then be aware of movement and packets could be optimally routed using the now visible destination IP address This has issues for the transport protocol … can SCTP and HIP help here? The application layer will also be aware of mobility IF it binds to the IP address With the industry trying to move away from apps binding themselves to IP addresses … what other options might exist to make the app aware of movement? What interesting things can we do if apps are aware of movement? 14) Exposing mobility to higher layers
15) The nature of paging • What about paging? • Mobile devices are severely power constrained and will drop into sleep mode at every opportunity. • How do we find and wake up these devices? • Do different apps require different treatment by the paging system? • Does paging apply to different layers?
16) Working with industry organizations • IEEE – we are strong here • IETF – we are strong here • 3GPP (3rd generation partners project) – we are almost totally invisible here. What do we need to be doing in 3GPP? How do we cover this group? What is our strategy here? Right now this organization is dominated by Ericson and NSN. • NGMN Forum (Next Generation Mobile Network) – Setup by major operators to try and push their agenda.
17) Privacy, security, and identity • How do we discover the networks we have roaming access to? • How is accounting across administrative boundaries handled? • How will the location information of the user be protected? • How are the user credentials protected in a roaming scenario? • How can we create a scalable roaming environment? • How do we prevent visited networks from creating false accounting data? • How do we authenticate the user for access to the network?