20
3G Application Aware RAN with In-bearer optimization Creating value from application prioritization Nokia Networks Nokia Networks white paper 3G Application Aware RAN with In-bearer optimization

Nokia 3g application_aware_ran_whitepaper

Embed Size (px)

Citation preview

Page 1: Nokia 3g application_aware_ran_whitepaper

3G Application Aware RAN with In-bearer optimization

Creating value from application prioritization

Nokia Networks

Nokia Networks white paper3G Application Aware RAN with In-bearer optimization

Page 2: Nokia 3g application_aware_ran_whitepaper

networks.nokia.comPage 2

Contents

   1.   Executive summary: Application differentiation creates added value in mobile broadband

3

2. Mobile data usage will continue to grow rapidly 4

3. Internet usage models shift rapidly 6

   4.   The evolution of end-to-end data traffic management 7

5. Vision of QoS versus operational practice 7

6. Business models 8

   7.   Existing approaches for enabling QoS differentiation per application

8

8. Application Aware RAN enables real application differentiation in 3G

10

   9.   Impact of 3G Application Aware RAN QoS on user QoE 11

10. The test cases explored 12

11. Application Aware RAN Test Results 13

11.1 Application Aware RAN priority greatly improves web browsing performance

13

11.2 Application Aware RAN priority boosts YouTube video performance

14

11.3 Application Aware RAN priority with in-bearer application optimization boosts application multi-tasking

15

12. End-to-end QoE measurement with performance manager and service quality manager for priority services

18

13. Find out more 18

14. Abbreviations 19

Page 3: Nokia 3g application_aware_ran_whitepaper

networks.nokia.comPage 3

1. Executive summary: Application differentiation creates added value in mobile broadband

Data usage is growing faster each year and application usage patterns are changing in unpredictable ways. Operators need better methods to cope with the dynamic data consumption unleashed by the presence of smartphones in existing 3G networks. These networks will continue to carry the majority of data traffic over the coming decade. Even as additional spectrum, small cell deployments, network features and optimization techniques help to increase capacity, users expect ever greater quality in their data sessions. Many subscribers consider data transmission quality to be as important as network coverage, voice quality, and price according to recent research1.

The ability to prioritize application traffic dynamically when needed or when it adds value for the user and the service provider in a simple but effective manner creates an opportunity for the operator to move up the data transport value chain beyond being a pipe provider. Existing industry solutions for creating application awareness have limitations that have hindered wide-scale adoption and cannot take advantage of information from the radio access network (RAN) about cell load and radio link conditions. In addition, existing solutions are complex to deploy and cannot react to real-time changes in network and application behavior.

Nokia Networks’ solution, 3G Application Aware RAN with in-bearer optimization, leverages existing core network capabilities to inspect data traffic at the application level while applying policy rules and enforcement in real-time and end-to-end. Nokia Networks combines Core Network intelligence with RAN awareness of cell load and radio link conditions at the bearer level to create a real-time solution for detecting application data and enforcing policy. Additionally, in-bearer application optimization takes service prioritization further by extending priority within the radio access bearer to treat applications with different latency requirements to assist multi-tasking application users.

The inclusion of the RAN to real-time QoS decisions is the missing link that gives operators real-time, intelligent control over applications, breaking all the operational limitations that previously prevented 3G networks from introducing application prioritization through transport differentiation. Now operators can create application-specific packages with personalization and targeted pricing to reflect measurable service quality.

1. http://networks.nokia.com/news-events/press-room/press-releases/mobile-operators-keep-your-customers-loyal-by-focusing-on-voice-data-quality-1gbperday

Page 4: Nokia 3g application_aware_ran_whitepaper

networks.nokia.comPage 4

Nokia Networks’ Smart Labs results show that per application-level detection is very effective and able to provide significant improvements in data throughput for prioritized services. Smart Labs conducted a series of tests in which diverse popular applications such as web browsing, YouTube, Skype, peer-to-peer (P2P) and file download operated on off-the-shelf Android devices while the cell load varied  from no congestion to high congestion. The prioritized application and user experienced the following benefits:

•  HTTP web browsing data throughput increased 1.65 times when prioritized in medium-loaded systems and 2.9 times in highly- loaded systems, compared with testing on a best effort basis in a congested cell.

•  Response times for web services were improved.

•  YouTube video data throughput increased 1.93 times when prioritized in medium-loaded systems and 2.7 times in highly-loaded systems compared with best effort carriage in a congested cell.

•  YouTube video stream re-buffering was reduced or eliminated with faster stream setup.

•  P2P traffic scheduling was more flexible.

Building on the capabilities of Application Aware RAN an additional feature for in-bearer application optimization enables service prioritization for multi-tasking application users within a radio access bearer for latency sensitive applications to be prioritized ahead of non-latency sensitive ones without changing the QoS Profile. 

Prioritized applications were found to experience the following benefits when tested on a cell with load versus when QoS was inactive:

•  FTP+HTTP Multi-tasking application throughput improved by 8 times

•  Web page download times were approximately 8 times faster.

•  FTP+YouTube Multi-tasking application throughput were 3 times faster

•  YouTube video streams did not suffer from re-buffering and buffering times were reduced from 52 seconds to 6 seconds.

Operators now have an effective system to offer per application priority at the subscriber level. It is operationally deployable, backwards-compatible to all 3G devices, and provides added, monetizable value for the priority delivery of data services.

2. Mobile data usage will continue to grow rapidly

The smartphone is driving continual increases of data consumption by subscribers on mobile networks with a five-fold increase in usage to 4 GB 

Page 5: Nokia 3g application_aware_ran_whitepaper

networks.nokia.comPage 5

per month by 2019 from .8 GB per month in 20142. Usage of web, video, audio and file sharing continues to rise due to a confluence of widespread availability of 3G networks, continued speed improvements from HSPA, and increases in smartphone penetration and device capabilities.

In 2013 mobile networks carried for the first time more than one exabyte (1 Billion Gigabytes) and the Cisco Visual Networking Index (VNI) projects that mobile networks will carry 2.5 exabytes of data per month and further predicts data traffic to exceed 24 exabytes per month by 2019 (Figure 1).

Mobile video which earlier in the decade first became the largest single traffic type on mobile networks continues to dominate traffic and is expected to grow to three-quarters of all mobile data traffic by 2019.

Demand for mobile data is closely correlated to the evolution of device and screen technologies, which are among the areas of the Information and Communication Technology (ICT) industry that are evolving the fastest. In 2007 the first iPhone® was introduced with a screen resolution of 320 x 480 pixels which in seven years increased by 13.5 times to a display containing 1920 x 1080 pixels in the iPhone 6 Plus which users are filling with content at two times the data usage of the “smaller” iPhone 63. Ultimately, only the human eye will limit the amount of digital content that will be consumed by a mobile device. In addition to consuming content, ubiquitous integrated cameras with high resolution and frame rate are producing exabytes of digital content to be distributed via networks.

Clearly, mobile networks are facing a growing possibility of congestion during peak usage hours, despite investments in additional base stations, advanced RF features, and other capacity improvements.

Fig. 1. Cisco VNI global mobile data traffic growth. 

Exab

ytes

/mon

th

Mobile Data Traffic Growth 2019 mobile networks predicted to carry more than 24 Exabytes per month

22001144 22001155 22001166 22001177 22001188 22001199 0

24

12

2. http://www.cisco.com/c/en/us/solutions/service-provider/visual-networking-index-vni/index.html 3. http://www.citrix.com/content/dam/citrix/en_us/documents/products-solutions/citrix-mobile-analytics-report-february-2015.pdf

Page 6: Nokia 3g application_aware_ran_whitepaper

networks.nokia.comPage 6

3. Internet usage models shift rapidlyWhile mobile internet network traffic continues to rise, there is a noticeable change in usage patterns. Video is embedding itself into more application categories and the types of application which subscribers use are changing to include new categories and at different times of the day.

For example according to a Citrix Mobile Analytics Report4, a new small but growing category, mobile dating, is used the most at 6 PM, while healthcare / fitness applications which grew from 39% to 78% of subscribers in two years have peak usage between 5-7 PM.

The usage of any content type is dynamic and in context of the application being used. Most YouTube users watch videos for less than 5 minutes at a time while on NetFlix a majority of users watch for more than 5 mins4. Embedding of video in applications has increased to include social media applications like Facebook to messaging applications like Snapchat / Instagram, and into new categories like mobile gaming where two years ago none of top five applications contained video to all five today4. Users and their devices are multi-tasking far more than before triggering multiple simultaneous data sessions with different QoS requirements.

Interestingly, according Nokia’s own Acquisition and Retention Study5 report 41% of customers expect excellent network quality even if it costs more. 3G networks need to be able to support rapid changes in usage which adjust to work at the speed of the user.

Fig. 2. Use of different smartphone applications by users.

Photos Video Music Games Shopping Productivity Storage

… … …

4. http://www.citrix.com/content/dam/citrix/en_us/documents/products-solutions/citrix-mobile-analytics-report-february-2015.pdf5. http://networks.nokia.com/news-events/press-room/press-releases/network-and-service-quality-keeps-customers-loyal-nokia-retention-study-shows

Page 7: Nokia 3g application_aware_ran_whitepaper

networks.nokia.comPage 7

Fig. 3. Current business models treat traffic equally under given conditions.

IMEI

Smartphone Tablet Thermostat

IMSI

Gold Silver Bronze

Volume

Monthly Daily

Time

9AM to 11 AM 4PM to 5PM

Location

Home Zone

Access

2G, 3G, LTE

4. The evolution of end-to-end data traffic management

How can operators and their customers adjust to the impact of data hungry applications, especially since many data plans have pre-defined usage limits and the popularity of different applications keeps changing? Traditional traffic management and billing models (Figure 3) are inflexible on a per application basis.

Most types of data traffic are treated equally under a given set of conditions, such as device type (IMEI), subscriber level identity (IMSI), access type (2G, 3G, LTE), time of day, data volume and location.

5. Vision of QoS versus operational practice

When the wireless industry standardized Quality of Service (QoS) differentiation mechanisms in the 3GPP (Third Generation Partnership Project) more than a decade ago, Access Point Names (APN) with separate primary bearer (Packet Data Protocol Context, of “PDP Context”) were created to support QoS for data connections. However, it proved impractical to manage multiple APNs per device across the network as the number of applications and connections proliferated. This challenge was compounded by the complexity of 3G QoS and associated device support, which meant that the QoS mechanism was not used to the fullest extent possible in operational networks.

The result of treating different traffic types equally or with limited number of levels owing to the limitations of PDP context-level QoS is a constrained business model. Single payer models, preferred and paid prioritization can’t bring as much value and flexibility either to the customer or the service provider because of the lack per application-level QoS differentiation.

If application-level differentiation can be enabled, then per application- level management of the Quality of Experience (QoE) is possible, enabling less important data to be delayed and preferred data to be prioritized. New application-level pricing models can be offered thanks to transport being a value added delivery service, rather than a best effort pipe.

Page 8: Nokia 3g application_aware_ran_whitepaper

networks.nokia.comPage 8

6. Business modelsClever, tiered pricing models and bundles containing mobile broadband data services are major tools to help operators combat revenue erosion. This will only become more important going forward as operators explore new models, such as price differentiation by quality and application.

Ultimately, however, there is an increasing danger that the business of being a network operator is changing from a retail model to a utility type of business, with limited room for positive differentiation. Handset vendors and OTT providers are gaining more traction with consumers and it will be harder than ever for operators to establish strong customer relationships in the future as consumers are focused on the latest device and the coolest app. In most markets, customer loyalty is in decline. Consumers are increasingly selecting their mobile broadband service provider based on coverage, performance along with service price and handset offers.

It’s also clear that the number of cooperation agreements between operators, OTT vendors and other industries will increase significantly, especially in the areas of content delivery. Delivering services in a differentiated and managed way opens up additional personalization and monetization opportunities in partnership with content providers and global content delivery networks (CDNs) by providing a clear value-add to the partners in the value chain and ultimately the end users. A good analogy would be a value-adding logistical service such as a premium postal service offering fast and reliable delivery.

Consider the example of video traffic delivery. The video traffic quality issue can only be rectified by the network operator. The operator owns and operates the only portion of the network between video servers and digital video players that does not carry an explicit Service Level Agreement (SLA). If operators can ensure a better service quality for specific OTT video streams and provide SLAs on those streams that include the journey through the RAN, various parties including consumers might be willing to pay for the value added transport. Content providers want their end users to receive their content at a reasonable quality. There are several potential revenue sources for the operator: the end user paying for “premium” internet TV, the global CDN, the content aggregator and the content provider paying for an explicit SLA (in markets which allow various forms of paid prioritization).

7. Existing approaches for enabling QoS differentiation per application

The industry has created a number of QoS differentiation solutions in an attempt to solve the need for application differentiation.

These solutions have all seen some level of adoption depending on the needs of the 3G HSPA network operator but they each have limitations that have prevented their use on a large scale.

Page 9: Nokia 3g application_aware_ran_whitepaper

networks.nokia.comPage 9

•  Core based application throttling

Application throttling is triggered by deep packet inspection (DPI) based application detection, subscription, fair use policy, time of day, the user’s initial cell location and a prediction of cell peak hours. Application IP flow throttling is enforced within the core network.

Limitations: The system is not aware of cell-level loading in real time. To add cell load awareness requires complex system integration (OSS, Policy, GW, DPI) and will be inefficient and inaccurate.

•  Network-requested PDP context QoS modification

  Trigger for modification of the PDP context bearer is same as in core based application throttling.

  Limitations: Modifications impact the whole bearer so all applications are affected by any change. Frequent modifications cause high signaling load in all network elements (GGSN, SGSN, RNC, NodeB).

•  Dedicated access point name (APN) per application

  Primary PDP context level QoS differentiation can be provided by an application-specific APN (device OS/application support), which is limited to certain services, domains and operator or partner content. Normally only operator service specific APNs are supported (e.g. MMS, IMS). Thus application specific APNs are not really an option.

Limitations: There is no true application awareness within the PDP context to determine which applications benefit. Typical usage in networks is limited to specific services with policy rules. APN configuration information requires the operator to push terminal configuration parameters to the device and provide support from device software. It’s operationally complex to implement, manage and maintain.

•  Network-requested secondary PDP context and dynamic application mapping

Selected applications are detected by DPI and secondary PDP context is established for application specific traffic. Traffic flow templates in core and user equipment (UE) map application IP flows to secondary PDP context in order to provide differential QoS.

Limitations: This is only supported by LTE terminals and not currently by 3G (or 2G) terminals. It creates challenges in handling a mass of short-lived uplink flows (such as P2P demotion). It creates delay because of the need to activate radio resources when the first data arrives at the dedicated bearer.

Each of these existing solutions solves some problems, but none of them fully address radio access, which is the best real-time enforcement point for per application QoS differentiation. Dynamic radio access scheduling must be combined with the core network’s control and logic enforcement in order to react dynamically to network conditions and user application usage.

Page 10: Nokia 3g application_aware_ran_whitepaper

networks.nokia.comPage 10

8. Application Aware RAN enables real application differentiation in 3G

The wireless industry needs an end to end, per application level QoS solution, but has not been able to implement a comprehensive system for 3G QoS.

Nokia Networks has innovated with the creation of 3G Application Aware RAN with in-bearer optimization an end to end QoS solution which works with all existing HSPA-capable devices allowing operators and their customers to prioritize important and specific data traffic flexibly, without operational complexity and past limitations.

Nokia Networks’ Application Aware RAN is a dynamic, real-time solution which (Figure 4) connects all the needed subsystems into one end to end chain for QoS differentiation down to a per subscriber per application level even within the same radio access bearer (RAB) which opens up the possibility to prioritize multi-tasking subscribers who may be using foreground and background applications simultaneously.

The solution works within the current 3GPP standards and network elements by using QoS policies from the Subscriber Profile Repository (SPR) which are used by the policy charging and rules function (PCRF) to program the policy control enforcement function (PCEF) to mark IP packets which are detected using deep packet inspection (DPI). These markings are used by the Radio Network Controller (RNC) for bearer priority and by the BTS to dynamically change bearer priorities in real-time using Nokia’s advanced radio scheduling algorithms. Fast radio scheduler reactions react to changing cell loads, radio conditions and policy needs.

3G BTS

RNC

Define application and subscriber specificQoS profiles

Internet

Charging

PCEF + DPI

> DPI: monitor and detect application usewhile marking applications according to policies

SGSN PCRFSPR

Supported byall devices

> Real-time QoS enforcement and cell load awareBest QoE and efficiency of the most criticalsystem resources

>OSS

Fig. 4. Nokia Application Aware RAN end-to-end system approach.

Page 11: Nokia 3g application_aware_ran_whitepaper

networks.nokia.comPage 11

Test UE

Other UE’s for Load

Application Aware RAN UserBest Effort or lower priority User

RNC

SGSN

PCEF

PCRF

3GBTS

Fig. 5. Lab setup for testing 3G Application Aware RAN and Application Aware RAN with in-bearer  application optimization

9. Impact of 3G Application Aware RAN QoS on user QoE

In order to benchmark the efficacy of the Nokia 3G Application Aware RAN solution, a series of lab tests were conducted in the Nokia Networks’ Smart Labs using a 3G HSPA network with commercially available HSPA Android-based smartphones. Note that Nokia’s Application Aware RAN solution is network-based, dynamic, cell-load aware and terminal-independent, so it supports all HSPA devices. Depending on the test, five different common smartphone activities were tested, including web browsing, file download, YouTube, P2P torrent, email and Skype with and without multi-tasking.

The test setup is shown in Figure 5. Note that real-world results may vary from lab

Page 12: Nokia 3g application_aware_ran_whitepaper

networks.nokia.comPage 12

10. The test cases exploredThe test labs looked at five types of applications commonly used in an HSPA network under varying levels of load. They tested the impact of per application priority setting versus no priority as a best effort application.

Additional scenarios explored real-world user behavior, which typically involves using multiple applications simultaneously on one device with priority-setting as the network load varies.

For basic Application Aware RAN, Nokia Smart labs tested the selective prioritization of applications on a device in preference to other applications on the same device. Once the base test cases were completed then Application Aware RAN with in-bearer application optimization was tested to find the performance when users multi-task and use FTP+HTTP or FTP+YouTube which varying priority needs. 

Table 1. Description of test cases to verify the impact of Application Aware RAN.

Test Scenario Test Description

Unloaded system Single application (web browsing, file download, YouTube, P2P torrent  and Skype) with PRIORITY

Medium cell load Single application with NO PRIORITY (best effort) 

Single application with priority

Application Aware RAN

High cell load  Single application with NO PRIORITY (best effort) 

Single application with priority

Application Aware RAN

in-bearer optimization FTP+HTTP

FTP+HTTP on single RAB with cell load NO PRIORTY (best effort) 

Multi-tasking with varying application priority

Combination of Application Aware RAN and in-bearer application optimization

in-bearer optimization FTP+YouTube

FTP+YouTube (HD video 720p)  on single RAB with cell load NO PRIORITY (best effort) 

Multi-tasking with varying application priority

Combination of Application Aware RAN and in-bearer application optimization

Page 13: Nokia 3g application_aware_ran_whitepaper

networks.nokia.comPage 13

11. Application Aware RAN Test Results 11.1 Application Aware RAN priority greatly

improves web browsing performanceNokia Networks’ Smart Labs testing of prioritized HTTP web browsing with Application Aware RAN shows significant performance improvements resulting in higher user satisfaction for prioritized versus non-prioritized (best effort) sessions during periods of congestion.

Test results showed (Figure 6) remarkably improved service quality for a user with web browsing priority under different load conditions:

•  Under medium cell load with prioritization, HTTP throughput increases from 4.1 Mbps to 6.8 Mbps or 1.65 times compared with tests with no priority as a best effort application.

•  Under high cell load with prioritization, HTTP throughput increases from 1 Mbps to 2.9 Mbps or 2.9 times compared with tests with no priority as a best effort application.

•  All tests show a general improvement in response times for web services.

10

2

4

6

8

12

0HTTP PriorityNo Load

HTTPPriorityMediumLoad

HTTP Web Browsing Results

No HTTPPriorityMediumLoad

No HTTPPriorityHigh Load

HTTPPriorityHigh Load

Dat

a Ra

te (M

bps)

Application Aware RAN Increases HTTPThroughput in Cell CongestionUser with HTTP priority maintains higher data rates

Fig. 6. HTTP browsing results, prioritized vs. non-prioritized.

Page 14: Nokia 3g application_aware_ran_whitepaper

networks.nokia.comPage 14

11.2 Application Aware RAN priority boosts YouTube video performance

If Application Aware RAN prioritization can improve web browsing for selected users during periods of congestion, what effect can it have on demanding video sessions? Nokia Networks’ Smart Labs applied application priority to a user streaming a 30 second, 720p YouTube video clip to an off-the-shelf Android device. Cell-level congestion conditions were varied from no load to high load using additional devices. A performance comparison of YouTube sessions with application priority and with no application priority (normal best effort data) was conducted.

Application Aware RAN created significant performance improvements in data throughput at times of congestion for a user with a prioritized YouTube service (Figure 7):

•  Under medium cell load with prioritization, throughput increases from 3.2 Mbps to 6.2 Mbps or 1.93 times compared with tests with no priority as a best effort application.

•  Under high cell load with prioritization, throughput increases from  1 Mbps to 2.7 Mbps or 2.7 times compared with tests with no priority as a best effort application.

Fig. 7. YouTube video session results, prioritized vs. non-prioritized.

5

1

2

3

4

6

7

8

0Video PriorityNo Load

VideoPriorityMediumLoad

YouTube Video Streaming Results

No VideoPriorityMediumLoad

No VideoPriorityHigh Load

VideoPriorityHigh Load

Dat

a Ra

te (M

bps)

Application Aware RAN Boosting video streaming performance.User with YouTube Priority has higher data rates for video sessions duringcongestion with faster server access and less buffering.

Page 15: Nokia 3g application_aware_ran_whitepaper

networks.nokia.comPage 15

•  The user with a prioritized YouTube service also experiences faster server access, making it quicker to set up a video stream. More importantly, when the cell experiences high load, video buffering times are substantially decreased.

•  YouTube data is successfully detected and prioritized, while other application data continues as best effort traffic.

11.3 Application Aware RAN priority with in-bearer application optimization boosts application multi-tasking

If Application Aware RAN prioritization alone can improve performance what is the impact of enabling the in-bearer application optimization feature along with Application Aware RAN?

In-bearer application optimization goes to the next level of Application Aware RAN by prioritizing applications or services within an access bearer where different application types can be scheduled according to their latency sensitivity. This reflects real-world smartphone scenarios with multiple applications running in parallel.

Nokia Networks’ Smart Labs applied both Application Aware RAN and in-bearer application optimization to observe the affect of both features running together when a smartphone is multi-tasking by operating two different types of applications simultaneously. 

For the tests, off the shelf Android devices were used on a loaded cell with the UE of interest being tested multi-tasking first FTP+ HTTP and then FTP+YouTube.

Application Aware RAN with in-bearer application optimization created significant performance improvements in data throughput and experience for the user and for the preferred application(s) versus if no QoS policies were active (Figure 8) in loaded conditions:

•  With priority for HTTP data traffic multi-tasking with a background application, application throughput increased from 1.07 to 2.31 Mbps for FTP and from 0.34 to 2.85 Mbps for HTTP compared to when QoS was inactive for an 8 times HTTP throughput improvement

•  With priority for HTTP data traffic multi-tasking with a background application, webpages load times were reduced from 141 to 18 seconds as compared to when QoS was inactive for almost a 8x reduction in time

Page 16: Nokia 3g application_aware_ran_whitepaper

networks.nokia.comPage 16

•  With priority for YouTube data traffic multi-tasking with a background application, throughput Increased from 0.6 to 2.8 Mbps for FTP and from 0.7 to 2.1 Mbps for YouTube compared to when no QoS was inactive for a 3 times YouTube throughput improvement

•  When YouTube is multi-tasking with a background application, initial video buffering times were reduced from 52 seconds with no QoS enabled to 6 seconds when both Application Aware RAN and in-bearer application optimization were operational and user annoying re-buffering events dropped from 59 to 0 events

0

1

2

3 UE FTP TP

UE HTTP TP

0.34

2.31

1.07

Dat

a Ra

te (M

bps)

No QoS Application Aware RAN + In-bearer App Optimization

FTP + HTTP

0

1

2

3 UE FTP TP UE YouTube TP

2.1

0.7

2.8

0.6

Dat

a Ra

te (M

bps)

No QoS Application Aware RAN + In-bearer App Optimization

FTP + YouTube

2.85

141 28 18 0

20

40

140

No QoS Application aware RAN Application aware RAN + In-bearer App Optimization

Web

pag

e do

wnl

oad

time

(sec

)

SSmmaarrttpphhoonnee mmuullttii--ttaasskkiinngg iinn llooaaddeedd cceellll..

BBaacckkggrroouunndd FFTTPP ++ ffoorreeggrroouunndd BBrroowwssiinngg..

Application Aware RAN with in-bearer application optimization improves both the application performance vs. no QoS and further increases the throughput and helps delay sensitive applications in cell load conditions

Application Aware RAN optimization improves user experience by reducing web page download times

Page 17: Nokia 3g application_aware_ran_whitepaper

networks.nokia.comPage 17

52 9 6 59 0 0 0

10

50

60

No QoS Application aware RAN

Application aware RAN + In-bearer App Optimization

SSmmaarrttpphhoonnee mmuullttii--ttaasskkiinngg iinn llooaaddeedd cceellll..

BBaacckkggrroouunndd FFTTPP ++ ffoorreeggrroouunndd YYoouuTTuubbee HHDD 772200pp..

YouTube initial buffering (sec)

Number of YouTube re-buffering events

Application Aware RAN optimization improves user experience by reducing video buffering times from click to view vs No QoS and reduces annoying re-buffering  of videos

Page 18: Nokia 3g application_aware_ran_whitepaper

networks.nokia.comPage 18

12. End-to-end QoE measurement with performance manager and service quality manager for priority services

Ensuring service quality for more demanding applications such as YouTube requires Operational Support Systems (OSS) to let the operator know that the enabled priority plan is working as expected.

Nokia Networks has designed OSS support to monitor application performance in management systems, with an overview and drill down support available for Application Aware RAN. Operators can monitor high-value applications within the network 24/7, with views of differentiated application throughput at the cell level and root cause analysis of service degradation, if it occurs. With the integrated measurement capability from Nokia Networks’ management service, operators can see QoE analysis with easy reporting of differential throughput for users and applications.

13. Find out moreContact Nokia Networks for more details and the results of the other Smart Labs tests for Application Aware RAN for 3G and how Nokia can help you add value from prioritization to your network.

Measure active application throughput for services for end user experienced DL throughput

Monitor application performance using Performance Manager/Service Quality Manager

> Holistic view or drill down service level> View differentiated app throughput

easily - per app priority class & cell

Performance Management for precise following of application priority classesKnow what you deliver to meet marketing promises & ensure great user experience

Fig. 9. Nokia performance management systems for application monitoring.

Page 19: Nokia 3g application_aware_ran_whitepaper

networks.nokia.comPage 19

14. Abbreviations2G  Second Generation cellular 3G   Third Generation cellular3GPP   Third Generation Partnership ProjectAPN Access point nameBTS Base Transceiver StationCDN Content delivery networksDPI Deep packet inspectionNB NodeBGBR   Guaranteed bit rate GGSN   Gateway GPRS support nodeGW  GatewayHSPA  High-speed packet accessHTTP   Hypertext Transfer ProtocolIMSI International Mobile Subscriber IdentityIMEI International Mobile Equipment IdentityLTE Long Term EvolutionNW  NetworkOTT Over-the-topOSS Operational support systemsPDP Packet Data ProtocolP2P Peer-to-peerPCEF Policy Control Enforcement FunctionPCRF Policy Charging and Rules FunctionQCI QoS Class IndicatorQoS Quality of serviceQoE Quality of experienceRAN Radio access networkRNC Radio Network ControllerSGSN   Serving GPRS Support NodeSLA Service Level AgreementUE User equipment

Page 20: Nokia 3g application_aware_ran_whitepaper

networks.nokia.comPage 20

Nokia is a registered trademark of Nokia Corporation. Other product and company names mentioned herein may be trademarks or trade names of their respective owners.

Nokia Nokia Solutions and Networks Oy P.O. Box 1 FI-02022 Finland

Visiting address: Karaportti 3, ESPOO, Finland Switchboard +358 71 400 4000

Product code C401-01176-WP-201503-1-EN

© Nokia Solutions and Networks 2015