8 Building the Mobile Internet 5,0001 ■Mobile VolP 4.500 ■Mobile Gaming ■Mobile P2P 4.000 Mobile Web/Data 3.500 ■Mobile Video -4% 3,000 8% 5% 2,500 17% 2.000 1,500 1.000- 66% 500 0 2009 2010 2011 201220132014 Figure 1-7 Global Mobile Data Traffic Growth by Content oeL uJajul al!qow 00:00 04:00 08:00 12:00 16:00 2000 Time Figure 1-8 Example Diurnal Variarion of Mobile Internet Traffic Load Note You can also use the information in Figure 1-8 to estimate how much of the daily traffic is consumed within the busy hour.With a throughput profile as shown,we can esti- mate that around 6 percent of daily mobile Internet consumption occurs in the busy hour. This compares with traditional telephony networks,which are typically dimensioned to support up to 20 percent of daily calls during the busy hour. The final characteristic of future mobile consumption will be its"always-on"nature. Historically,while cellular network standards have had the capability to support always- on users,there were often implementation restrictions with the number of always-on users a system could support.So,cellular standards allowed mobile data-enabled devices to be attached to a cellular network without allocating them an IP address.Equally,lega- cy cellular networks will be typically configured to automatically deallocate a device's IP From <www.wowebook.com>
ptg Figure 1-7 Global Mobile Data Traffic Growth by Content 8 Building the Mobile Internet 5,000 4,500 4,000 3,500 3,000 2,500 2009 Mobile VoIP Mobile Gaming Mobile P2P Mobile Web/Data Mobile Video 2010 2011 2012 2013 2014 2,000 1,500 1,000 500 0 8% 17% 66% 4% 5% PB/mo 00:00 04:00 08:00 12:00 16:00 20:00 Time Mobile Internet Traffic Figure 1-8 Example Diurnal Variation of Mobile Internet Traffic Load Note You can also use the information in Figure 1-8 to estimate how much of the daily traffic is consumed within the busy hour. With a throughput profile as shown, we can estimate that around 6 percent of daily mobile Internet consumption occurs in the busy hour. This compares with traditional telephony networks, which are typically dimensioned to support up to 20 percent of daily calls during the busy hour. The final characteristic of future mobile consumption will be its “always-on” nature. Historically, while cellular network standards have had the capability to support alwayson users, there were often implementation restrictions with the number of always-on users a system could support. So, cellular standards allowed mobile data–enabled devices to be attached to a cellular network without allocating them an IP address. Equally, legacy cellular networks will be typically configured to automatically deallocate a device’s IP From <www.wowebook.com>
Chapter 1:Introduction to "Mobility"9 address after a period of inactivity.This can be contrasted with wireless LAN networks, which have been based on the assumption of an always-on Ethernet service.However,the next generation of cellular standards has been designed to support only always-on behav- ior.For example,whenever a device attaches to an all-IP LTE network,it will,by default, receive an IP address and be automatically enabled to send and receive IP packets. Mobile Challenges With the massive increase in forecasted consumption of mobile Internet services,it is interesting to understand how today's macro-cellular networks are positioned to accom- modate this unparalleled growth in capacity.A simple way to evaluate the capacity of a conventional cellular system is to examine three key cellular characteristics: ■ Spectrum:The amount of available radio frequency spectrum is a key factor in determining the capacity of a mobile system.The spectrum suitable for mobile net- works is a scarce resource.It must be available in sufficient bandwidths to support higher-speed access while also providing good propagation characteristics required for providing wide-area coverage. Spectral efficiency:The efficiency by which spectrum is used is another critical fac- tor in determining the capacity of a mobile system.There is an upper bound to the amount of information that can be transferred in a given bandwidth which is subject to background noise,termed the Shannon Limit.It is generally accepted that as more advanced signal-processing techniques are applied to mobile communications systems,they are rapidly approaching such a theoretical limit. Frequency reuse:Because of its relative scarcity,mobile systems are required to reuse their allocated radio spectrum across a particular network of cell sites. Increasing the capacity by frequency reuse typically means dividing an existing cell into multiple smaller cells. Agilent has analyzed the growth in average macro-cellular spectrum efficiency as cellular technologies have evolved from the original GSM system to the latest LTE-A standards'. In parallel,the growth in available spectrum for cellular systems can be mapped.Figure 1-9 shows both of these growth lines over the period from 1990 to 2015(with efficiency and spectrum being normalized to 1992 levels).The figure also shows the product of the two variables,indicating the estimated growth in average macro-cellular system capacity. Whereas the Cisco VNI forecast estimates that demand for mobile Internet traffic will grow 39-fold over a five-year period,Figure 1-9 illustrates that by only using better radios and more spectrum,the average capacity gains achievable by adopting a conven- tional macro-cellular approach will see mobile broadband capacity a supply increase by a factor of fourfold over the same five-year period.The only way that networks are going to be able to support the estimated increase in mobile broadband traffic is by increasing the adoption of smaller cells,with the chart in Figure 1-9 forecasting that the number of cells will need to increase tenfold to meet the expected demand in traffic. From <www.wowebook.com>
ptg address after a period of inactivity. This can be contrasted with wireless LAN networks, which have been based on the assumption of an always-on Ethernet service. However, the next generation of cellular standards has been designed to support only always-on behavior. For example, whenever a device attaches to an all-IP LTE network, it will, by default, receive an IP address and be automatically enabled to send and receive IP packets. Mobile Challenges With the massive increase in forecasted consumption of mobile Internet services, it is interesting to understand how today’s macro-cellular networks are positioned to accommodate this unparalleled growth in capacity. A simple way to evaluate the capacity of a conventional cellular system is to examine three key cellular characteristics: ■ Spectrum: The amount of available radio frequency spectrum is a key factor in determining the capacity of a mobile system. The spectrum suitable for mobile networks is a scarce resource. It must be available in sufficient bandwidths to support higher-speed access while also providing good propagation characteristics required for providing wide-area coverage. ■ Spectral efficiency: The efficiency by which spectrum is used is another critical factor in determining the capacity of a mobile system. There is an upper bound to the amount of information that can be transferred in a given bandwidth which is subject to background noise, termed the Shannon Limit. It is generally accepted that as more advanced signal-processing techniques are applied to mobile communications systems, they are rapidly approaching such a theoretical limit. ■ Frequency reuse: Because of its relative scarcity, mobile systems are required to reuse their allocated radio spectrum across a particular network of cell sites. Increasing the capacity by frequency reuse typically means dividing an existing cell into multiple smaller cells. Agilent has analyzed the growth in average macro-cellular spectrum efficiency as cellular technologies have evolved from the original GSM system to the latest LTE-A standards8 . In parallel, the growth in available spectrum for cellular systems can be mapped. Figure 1-9 shows both of these growth lines over the period from 1990 to 2015 (with efficiency and spectrum being normalized to 1992 levels). The figure also shows the product of the two variables, indicating the estimated growth in average macro-cellular system capacity. Whereas the Cisco VNI forecast estimates that demand for mobile Internet traffic will grow 39-fold over a five-year period, Figure 1-9 illustrates that by only using better radios and more spectrum, the average capacity gains achievable by adopting a conventional macro-cellular approach will see mobile broadband capacity a supply increase by a factor of fourfold over the same five-year period. The only way that networks are going to be able to support the estimated increase in mobile broadband traffic is by increasing the adoption of smaller cells, with the chart in Figure 1-9 forecasting that the number of cells will need to increase tenfold to meet the expected demand in traffic. Chapter 1: Introduction to “Mobility” 9 From <www.wowebook.com>
10 Building the Mobile Interne 1000 Macro Capacity s100 Macro-Cel Efficiency 0 Spectrum 1990 1995 2000 2005 2010 2015* Figure 1-9 Estimating Future Average Macro-Cellular System Capacity So.w ecan predict that buiding he fuure mobile nteret working technology that following Scalable adoption of small-cell technologies-for example,using unlicensed IEEE 802.11 technology or licensed cellular-based home base station solutions A massive number of always-on devices,including scenarios where single subscribers have access to multiple devices Ubiquitous access,including nomadic access from in buildings as well as wide-area mobile access for on-the-go consumption Seamless access to a range of mobile services,including video,web access,peer-to- peer,Voice ver IP (VolP),and gaming services All of this would be easily achievable if the Internet supported native mobility. Unfortunately,the Internet is not mobile!Conventional approaches for delivering mobility have been to layer tunnels on top of the native Internet Protocol (IP).While casual observers might comment that wide-area mobility has been solved by the mobile broad- band a archi itectures developed by the cellular organizati d by or even the latest all-IP e networks archite he WiMAX forum,these all prov vide mobi lity function ality by tunneling packets over the Internet GPRS Tunneling Protocol(GTP):An overlay IP-based mobility protocol defined by 3GPP to provide mobility services for accessing the Internet by GPRS,WCDMA, and LTE-based radio access networks. Mobile IP(MIP):An overlay mobility protocol defined by the Internet Engineering Task Force(IETF)to provide IP mobility services and adopted by 3GPP2 to provide mobility services for accessing the Internet by cdma2000-based radio access net- works. From <www.wowebook.com>
ptg Figure 1-9 Estimating Future Average Macro-Cellular System Capacity So, we can confidently predict that building the future mobile Internet will require networking technology that supports the following: ■ Scalable adoption of small-cell technologies—for example, using unlicensed IEEE 802.11 technology or licensed cellular-based home base station solutions ■ A massive number of always-on devices, including scenarios where single subscribers have access to multiple devices ■ Ubiquitous access, including nomadic access from in buildings as well as wide-area mobile access for on-the-go consumption ■ Seamless access to a range of mobile services, including video, web access, peer-topeer, Voice over IP (VoIP), and gaming services All of this would be easily achievable if the Internet supported native mobility. Unfortunately, the Internet is not mobile! Conventional approaches for delivering mobility have been to layer tunnels on top of the native Internet Protocol (IP). While casual observers might comment that wide-area mobility has been solved by the mobile broadband architectures developed by the cellular organizations, or even the latest all-IP mobile networks architected by the WiMAX forum, these all provide mobility functionality by tunneling packets over the Internet: ■ GPRS Tunneling Protocol (GTP): An overlay IP-based mobility protocol defined by 3GPP to provide mobility services for accessing the Internet by GPRS, WCDMA, and LTE-based radio access networks. ■ Mobile IP (MIP): An overlay mobility protocol defined by the Internet Engineering Task Force (IETF) to provide IP mobility services and adopted by 3GPP2 to provide mobility services for accessing the Internet by cdma2000-based radio access networks. 10 Building the Mobile Internet 1000 100 10 1 1990 1995 2000 2005 2010 2015 Growth Spectrum Average Macro-Cell Efficiency Macro Capacity From <www.wowebook.com>
Chapter 1:Introduction to"Mobility"11 Control and Provisioning of Access Points(CAPWAP):An overlay IP-based mobili ty protoc col defined by the IETF to provide mobility services for accessing the Internet by IEEE 802.11-based radio access networks. Proxy Mobile IP (PMIP):An overlay IP-based mobility protocol defined by the IETF to provide mobility services that have been adopted by those architectures for accessing the Internet by IEEE 802.16e-based WiMAX radio access networks. As you will see later,Mobile IPv6 does have an optimized routed mode whereby,after having originally traversed a tunnel,packets are subsequently routed directly between a mobile node and a correspondent.However,you will also see how the use of IPv6 rout- ing headers and destination options in the operation of optimal routing simply moves the tunnel operation into the IP host;the IP applications running on a mobile node still need to be shielded from the operation of MIPv6 with route optimization. Quite clearly,with the number of mobile broadband users set to eclipse the number of fixed users and with ABI Research estimating that 1.15 billion handsets were sold in 2009,"prob- lems of mobility are upon us."The industry will soon be faced with the situation that the default technique for accessing the Internet will be through a mobility tunnel.Because tun- neling of traffic requires stateful tunnel gateways to be operated,services accessed through the"mobile Internet"might end up being more brittle than those accessed through the native IP networks on top of which the mobility tunnels are transported. Summary The next decade will see the convergence of mobility and the Internet.The various SDOs that define the next ge 2l-Pp”vision,pro neration of wide-area mobile e works have all aligned around an ccess to pu y IP-based packe ervices,with networks that can be transported over converged packet-based networks However,we confidently predict that the mobility use cases will broaden from today's homogeneous,cellular-only view of the world Devices will become more heterogeneous from an access perspective.Wi-Fi dual- mode capabilities will become widely integrated into the next generation of cellular devices Users will increasingly have access to more than a single device for accessing the mobile Internet. etworks will have to support an increasing amount of con sumption fro ocations,requiring techr sto be able to cost-effectively integrate small-cell technologies into the overall mobil These transitions are triggering a reassessment of how best to build the mobile Internet. for providing mobility inna-Pworld. From<www.wowebook.com>
ptg ■ Control and Provisioning of Access Points (CAPWAP): An overlay IP-based mobility protocol defined by the IETF to provide mobility services for accessing the Internet by IEEE 802.11–based radio access networks. ■ Proxy Mobile IP (PMIP): An overlay IP-based mobility protocol defined by the IETF to provide mobility services that have been adopted by those architectures for accessing the Internet by IEEE 802.16e–based WiMAX radio access networks. As you will see later, Mobile IPv6 does have an optimized routed mode whereby, after having originally traversed a tunnel, packets are subsequently routed directly between a mobile node and a correspondent. However, you will also see how the use of IPv6 routing headers and destination options in the operation of optimal routing simply moves the tunnel operation into the IP host; the IP applications running on a mobile node still need to be shielded from the operation of MIPv6 with route optimization. All this complexity is caused because, when IP was first proposed in 1975, “problems of . . . mobility were many years off.”9 Quite clearly, with the number of mobile broadband users set to eclipse the number of fixed users and with ABI Research10 estimating that 1.15 billion handsets were sold in 2009, “problems of mobility are upon us.” The industry will soon be faced with the situation that the default technique for accessing the Internet will be through a mobility tunnel. Because tunneling of traffic requires stateful tunnel gateways to be operated, services accessed through the “mobile Internet” might end up being more brittle than those accessed through the native IP networks on top of which the mobility tunnels are transported. Summary The next decade will see the convergence of mobility and the Internet. The various SDOs that define the next generation of wide-area mobile networks have all aligned around an “all-IP” vision, providing access to purely IP-based packet services, with networks that can be transported over converged packet-based networks. However, we confidently predict that the mobility use cases will broaden from today’s homogeneous, cellular-only view of the world: ■ Devices will become more heterogeneous from an access perspective. Wi-Fi dualmode capabilities will become widely integrated into the next generation of cellular devices. ■ Users will increasingly have access to more than a single device for accessing the mobile Internet. ■ Next-generation mobile networks will have to support an increasing amount of consumption from indoor locations, requiring techniques to be able to cost-effectively integrate small-cell technologies into the overall mobile network. These transitions are triggering a reassessment of how best to build the mobile Internet, comparing alternative techniques for providing mobility in an all-IP world. Chapter 1: Introduction to “Mobility” 11 From <www.wowebook.com>
12 Building the Mobile Internet Endnotes 1.http://www.mobilemarketer.com/cms/news/content/1877.html 2.http://www.morganstanley.com/institutional/techresearch/pdfs/ Internet_Trends_041210.pdf. 3."Mobile Devices Market Forecast Analysis,"ABI Research,June 2010. 4.http://www.oecd.org/dataoecd/19/40/34082594.xls. 5."The US Consumer PC Market In 2015."Forrester Research.June 2010. 6.http://www.cisco.com/web/about/ac79/docs/CLMW_Mobile_Internet_v20_ 072809FINALpdf. 7http://www.cisco.com/en/US/netsol/ns827/networking_solutions_sub_solution.html 8.M.Rumney,"IMT-Advanced:4G Wireless Takes Shape in an Olympic Year,"Agilent Measurement Journal,Issue 6.Sept.2008. 9.J.Day,Parterns in Nerwork Arcbirecture,Indianapolis,Indiana:Pearson Education; 2008. 10."Mobile Device Model Tracker,"ABI Research,April 2010. From <www.wowebook.com>
ptg Endnotes 1. http://www.mobilemarketer.com/cms/news/content/1877.html. 2. http://www.morganstanley.com/institutional/techresearch/pdfs/ Internet_Trends_041210.pdf. 3. “Mobile Devices Market Forecast Analysis,” ABI Research, June 2010. 4. http://www.oecd.org/dataoecd/19/40/34082594.xls. 5. “The US Consumer PC Market In 2015,” Forrester Research, June 2010. 6. http://www.cisco.com/web/about/ac79/docs/CLMW_Mobile_Internet_v20_ 072809FINAL.pdf. 7. http://www.cisco.com/en/US/netsol/ns827/networking_solutions_sub_solution.html. 8. M. Rumney, “IMT-Advanced: 4G Wireless Takes Shape in an Olympic Year,” Agilent Measurement Journal, Issue 6, Sept. 2008. 9. J. Day, Patterns in Network Architecture, Indianapolis, Indiana: Pearson Education; 2008. 10. “Mobile Device Model Tracker,” ABI Research, April 2010. 12 Building the Mobile Internet From <www.wowebook.com>