E2E Transport Network Solution

Advisory Report

Ericsson Demonstrates End to End Mobile Transport Solutions March 15, 2011  Issue

Mobile operators are being challenged to deliver on new high value services that stress their current mobile network infrastructures by requiring nearly unlimited levels of bandwidth, flexibility and to provide a high level of quality of experience; all the while lowering the price per bit which tends to limit service provider profit margins. In addition to economic and technical challenges, operators face the threat of subscriber churn if they do not deliver on expected services in a timely manner. Glen Hunt Current Anal ysi s Current Anal Principal Analyst, Transport and Routing Infrastructure

Ericsson’s end-to-end network infrastructure products offer operators, both fixed and mobile, a range of solutions that claim to address the most pressing issues as they continue their infrastructure investments. It is the “claims” “claims” portion of this proposition that the recent Ericsson mobile mobile transport infrastructure demonstrations demonstrations set out to prove. As part of this process, Ericsson and Current Analysis Analysis contracted Isocore to perform due diligence on the test plans proposed for a demonstration of Ericsson End-to-End mobile solution offering. Isocore provided comments and ensured through witnessing that the testing adhered to the final test plans. However, However, Isocore did not lead an in-depth testing activity and analysis of results. In February 2011, representatives representatives of Current Analysis Analysis and Isocore were given the opportunity opportunity to observe, first hand, hand, the finer points of delivering delivering a multigenerational multigenerational mobile network solution that addresses key pain points expressed by mobile operators as they sort through the myriad of options, architectures, and product/solution choices available to them in the marketplace. Te solution demonstrations were conducted in Ericsson’s Eurolab IP Networks facility located in Aachen Germany. Germany. High level tests were were executed that proved proved that the company’ company’s end to end mobile infrastructure infrastr ucture solutions soluti ons could deliver support suppor t for multiple generations generatio ns of mobile traffic, 2G, 3G and 4G/LE, 4G/L E, provide network synchronization, demonstrate the quality and service and resiliency of its microwavee transport solution, and show the solutions delivered the expected quality of experience microwav under simulated network anomalies. anomalies. In all cases, the test scenarios observed by the Current Analysis team members delivered the expected results – demonstrating the validity of Ericsson’s claims to deliver pre-tested and proven proven solutions for mobile backhaul, IP RAN and mobile core segments of the network. Te following sections of this report provide background regarding the requirements placed on infrastructure to meet expectations and describe the specific tests and observed results.  Current Perspective

The Challenge

Mobile operators face multiple challenges to support basic and advanced mobile services, which are transitioning from predominately predominately voice and limited data to a service mix which is trending toward

© 2011 Current Analysis Inc. All rights reserved. For more information, please call +1 703 404 9200, toll-free +1 877 787 8947 Europe +33 (0) 1 41 14 83 15. Or visit our Web site: www.currentanalysis.com www.currentanalysis.com

1

Advisory Report Report:

Ericsson Demonstrates End to End Mobile Transport Solutions Transport and Routing Infrastructure

an unpredictable mix of video, high bandwidth data services, plus the continued need to support existing 2G and 3G traffic. Complicating the picture are demands from mobile users who want to pay less per bit, expect always-on operation, and demands for a high quality of experience from their providers. Beyond cell site capacity issues which often gain center stage, the backhaul network has been and remains a well-known choke point in many networks. Operators evaluating backhaul solutions must also be cognizant of the interdependencies that RAN has with backhaul which includes how the multiple generations of traffic are supported, native or emulated transport, network timing and synchronization, network delay and jitter, and last by not least, resiliency and recovery from network failures – either equipment of circuit induced. Although the challenges are many, when viewed as an end to end transport solution, operators have a framework to address their specific needs. The Trends

Mobile operators can try, and in some cases, skip the complexity of supporting legacy services, and will often find the concept of building an all-IP infrastructure optimized to support 4G LE services attractive. Te all-IP model brings with it fewer “moving parts” and a true focus on delivering high bandwidth video and data services. However, reality quickly steps in and provides a level set – current services need to be supported, which brings the need to support traditional DM and AM-based support back into the mix. Mobile operators are also often forced to leverage existing infrastructure investments until they are fully depreciated or taken out of service. And, as hyped as 4G services appear in the media, the bulk of operator revenues come from existing services and from the continued evolution based on technologies such as HSPA+ and EVDO, which address an immediate need to provide higher performance 3G data services. Tere appears to be little doubt that the goal is to provide 4G services over an infrastructure that will represent a convergence of previous generations, with LE service materializing beginning in earnest in the 2011, the industry appears to be in agreement. The Solutions

Given the complexity of delivering mobile services, operators are challenged to make investment decisions, but often lack the assurances from their suppliers that these investments will position them to best address the needs of a highly flexible and unpredictable market place. Operators seeking solutions can take several approaches, they can follow the traditional route of selecting individual network components, such as cell site and aggregation routers, edge switches, mobile core, microwave, fiber access, and a management system to provide the operation and service provisioning support to deliver, trouble shoot and monitor network performance. An alternative approach where the operator leverages pre-integrated and pre-tested solutions appears to gaining momentum. Te solutions approach can one more viable and one that many operators are pursuing to reduce risk, improve time to market, optimize cost, and to obtain operational systems capable of enabling them to offer flexible high value ser vices, and position their infrastructures to meet the current and future challenges. Vendors such as Ericsson have packaged a combination of mobile transport solutions that enables operators to proceed with a high level of certainty and leverage an end to end solution that addresses the challenges noted above. Depending upon mobile operator preferences, they can leverage one of three solutions offered by Ericsson – or combine them to provide an end to end mobile transport network. Te three solutions separately include the Mobile Packet Backbone Network (MPBN), which provides a transport solution for 2G, 3G and LE mobile core networks; the IP RAN, which provides infrastructure and connectivity of GSM, WCDMA and LE radio networks; and the Mobile Backhaul (MBH), which provides the infrastructure to interconnect the IP RAN and core. Benefits of a solutions approach

Pretested and integrated solutions for the entire mobile network infrastructure, or segments of the infrastructure, can pay significant dividends for mobile operators in several areas, such as reduced © 2011 Current Analysis Inc. All rights reserved. For more information, please call +1 703 404 9200, toll-free +1 877 787 8947 Europe +33 (0) 1 41 14 83 15. Or visit our Web site: www.currentanalysis.com

2



testing and validation time, interoperability among the various components of the soltuion, proven backup and resiliency mechnaizms with known performance, management and provisioning support to reduce OpEx, and a consistent implementation of traffic prioritization and classification techniques to address the quality of experience so important to the end user and in the case of M2M traffi c the application. As an industry, there have seen multiple demonstrations of multi-vendor interoperability, including, Ethernet centric platforms to deliver cost effective high bandwidth services, emulation services for DM, rigorous timing and synchronization tests to prove field readiness of technologies such as IEEE 1588 and Synchronous Ethernet, and a few attempts to demonstrate end to end OAM and network performance monitoring. Operators; however, are becoming a bit more concerned about the ability of devices to work in concert to meet their overall services needs, less about the individual device specifications, and the vendor’s unit-level testing of those devices. Significant benefits can be realized by offering operators a fully tested solution that provides the scalability and flexibility to adapt to all forms of traffi c and service demands. Mobile End to End Transport Solutions

Current Analysis participated in suite of end to end mobile transport demonstrations based on Ericsson’s mobile transport infrastructure solutions. Te demonstrations were conducted in its Eurolab IP Networks facility located in Aachen Germany in February 2011. Te Ericsson team extracted several high level test scenarios that went toward proving the value in leveraging Ericsson’s pre-tested solutions and to show the depth of their transport solutions. esting included its optical, Ethernet and microwave products which together forms a complete network. Te demonstrations selected, verified the ability of the solutions to carry multiple generations of mobile traffi c, including LE. And, on the more subtle side, the demonstrations proved that the solutions could provide end to end quality of services based on traffic and subscriber profiles – a significant challenge facing most operators as they seek acceptable ways to deliver premium as well as best effort services. It is important to note that the demonstrations were not focused toward proving the functionality of each individual component or device in the transport network. Te following test cases were executed for the Current Analysis participants, and fully addressed the challenges noted above. For each case, we have noted the results observed from running the respective tests scenarios. In addition to observing the results of the tests, we were able to participate in creating network anomalies – some as severe as disrupting equipment power to demonstrate complete node failover and recovery. Perhaps the dream of every tester is to find a test scenario where a defect or un-explained behavior is detected, unfortunately, this did not occur in the testing, all of the test cases worked as designed and demonstrated advanced functionality based on a solid foundation. Test Environment

wo test profiles were used in the demonstrations, a baseline profile to create an artificial network load on top of the traffic created by the user applications, and to perform exact outage calculations during resilience tests. For that purpose, bidirectional traffic streams were injected into the network. A high load setup was also used for selected Quality of Service and load tests. Te streams defined were used to put additional traffi c on selected hops of the network in order to cause high load conditions on selected links/interfaces to induce overload, forced traffi c management to curb best effort traffic, and to show resiliency and recovery operations. Click here f or access to the test plan including details regarding the specifics of the test setup such as equipment type, interfaces, and configurations. 2G/3G End to End Integration

Tis test scenario was comprised of voice calls, internet access, video streaming, IPv6 transport, 2G and 3G mobile access. Te initial tests required that separate 2G and 3G voice calls be placed over the test infrastructure, which leveraged a remote site (Linköping, Sweden). Te test network © 2011 Current Analysis Inc. All rights reserved. For more information, please call +1 703 404 9200, toll-free +1 877 787 8947 Europe +33 (0) 1 41 14 83 15. Or visit our Web site: www.currentanalysis.com

3



was running the baseline traffic profile, and we where we were able to observe the video displaying an HD movie, best effort internet traffi c and video monitors showing the local site and the remote sites for the voice connection. We observed that the voice quality was good for both calls by using the appropriate hand sets and the applications running did not show any distress such as pixilation or distortion as a result of activating the voice calls. A 2G (GSM)/ 3G (UMS) radio bearer, respectively, was used on both sides involving the Ericsson BS, BSC, MSC-BC, M-MGw and HLR radio and core nodes. Te signaling and voice transport for the radio access of MS1/UE1 (Aachen) is accomplished using the MINI-LINK, SEA and OMS systems which are part of the Ericsson Mobile Backhaul (MBH) network solution build up in Aachen. In the Core Network, the call is routed between two Primary Sites of MPBN located in Aachen. o reach the radio access of MS2/UE2 (Linköping), the call was routed through the Ericsson Data Network to the remote cell site in Linköping. 3G PS Connection

Tis test was conducted using an Apple iPad to display a video using HP Progressive Download. A PC with 3G modem was connected to the test lab internal web server, which initiated pings continuously and downloads of large files using the ftp protocol. For each user device (UE), a 3G (HSPA) radio bearer was used involving the Ericsson RBS, RNC, SGSN-MME, GGSNMPG, SAPC and HLR radio and core nodes. Te signaling and data transport for the radio access of the UE was accomplished using the Ericsson Mobile Backhaul (MBH) network solution in Aachen. In the Core Network, the call was then routed between two Primary Sites of MPBN located in Aachen. Te connection to the Internet and the local web server was provided by a Juniper SRX Firewall/NA solution. Best effort quality of service was assigned to the established PDP contexts. Te baseline network profile was again used and the connections for the iPhone and PC/3G modem were connected, the video stream on the iPad was noted to have good visible quality (quality as observed), the ftp downloads were continuously running and reported not errors. We were able to observe bandwidth utilization on the Spirent test center, utilization tracked per the applications being executed. Since both, ftp download and video stream were transferred with best effort service, the video stream suffered from the high bandwidth demand of the ftp service unless the PC subscription is limited in bandwidth. Of significance with this test, is the treatment of best effort traffi c, which is allowed to consume available bandwidth which is not allocated to premium level traffi c, but there can be conflicts within the best effort where one application can interfere with others causing a level of service distress. Te test was repeated, and the video iPhone subscription was assigned as a premium 3G service. Te results showed that the stream, this time did not experience any distress due to the best effort traffi c, indicating that the QoS mechanisms were indeed operational. 3G PS Connection

A PC with 3G modem was connected to the test lab internal web server using the IPv6 protocol, as before the test initiates pings continuously and displays a web page. Te test verified that a true IPv6 connection was established and the ping recorded the round trip time over the connection, of significance was the observation that the round trip delay when compared to an IPv4 ping, were similar in time duration. It would have been somewhat expected that the process would have taken additional time due to the more extensive address lookup and browser processes. LTE End to End Integration

Tis test shows the integration of additional end to end use cases using LE mobile access technology, such as Internet access and video streaming, into the test network. Te testing verified that LE use cases are operational within the test network and shows that the network © 2011 Current Analysis Inc. All rights reserved. For more information, please call +1 703 404 9200, toll-free +1 877 787 8947 Europe +33 (0) 1 41 14 83 15. Or visit our Web site: www.currentanalysis.com

4



handles the quality of service and bandwidth requirements properly. Te cases included www and ftp services and HDV video streaming. An LE client and a live V session were demonstrated. Of particular importance was the ability to transport best effort traffi c without impacting the guaranteed traffic, indicating that the network properly prioritized and managed traffic flows. Te LE client connects to the Ericsson Evolved Packet Core via LE radio. Te network configuration was the same as in the previous test cases except that different protocols and links were used and the Ericsson CPG provided the SGi interface towards Internet and intranet servers. Best effort quality of service was assigned to the established evolved packet system bearers. Te test included pinging a node on the network, HP browsing and FP downloads. In addition, the LE client presented High Definition streaming videos from a local HP/streaming server or the Internet. Full LE throughput is shown end to end. Te basic testing confirmed that an LE subscriber could be provisioned, the applicable connections were made and a continuous ping could be setup, further a large file was successfully downloaded using Filezilla. Using best effort traffi c classification, a HD video was played, it was observed that best effort traffi c from the background traffi c could impact the perceived quality of the video. Another session was activated, this time with premium quality of service assigned, the results were similar to the 3G premium testing, where there the quality of experience was preserved and no degradation in the video steam was visible. Network Synchronization

Te test case verified that network synchronization functions performed reliably when delay and jitter are varied or when the bandwidth is reduced temporarily on links that carry synchronization signaling. Te base station was configured with two NP (Network ime Protocol) sources, when one of the sources was impaired, the base station switched to the alternative source. Te test was repeated using a PP (Precision ime Protocol) source, and when the primary source was impaired, the Site Integration Unit (SIU), which has a high stable oscillator (HS-OCXO), provided an accurate frequency that can be used as a timing source for base station (RBS) synchronization functionality (according to Ericsson- for several months), the eNodeB remained synchronized during this test.(Te test participants did not observe the test for months, but were assured that the high precision clocking in the SIU would indeed function as noted). Transport Microwave and Optical

Multiple transport tests were conducted using the MINI-LINK microwave and the OMS 1410 optical platforms. With the unquestionable trends toward adding capacity and resiliency to the mobile infrastructure, both solutions are needed to address these requirements. Te microwave solutions provide the first mile backhaul to service the increasing demands from the cell site; the optical solutions provide the same service when fiber connections are available. Of significance, is the ability of each platform to provide rigorous and consistent traffic management. Te transport tests noted below, and the QoE section that follows, rely heavily on these QoS and traffi c management mechanisms found in the two products. Common to both microwave and optical platforms is the ability to transport DM and packet traffic in their native format, the first test executed demonstrated this capability on each platform, and we noted that both streams were forwarded without any frame loss. Transport Microwave

Te microwave test scenarios demonstrated the reliability and quality of service capabilities of the MINI-LINK N transport solution, and leveraged technologies such as adaptive modulation, queuing and radio link bonding to deliver quality ser vices under impaired link conditions. Ericsson’s radio link bonding technique was shown, which when configured in a 2+0 configuration delivered a 100% increase in link performance, increased performance is critical to keeping up with bandwidth and subscriber growth. Since microwave transport represents a significant

© 2011 Current Analysis Inc. All rights reserved. For more information, please call +1 703 404 9200, toll-free +1 877 787 8947 Europe +33 (0) 1 41 14 83 15. Or visit our Web site: www.currentanalysis.com

5



alternative to fiber, it is imperative that the system deliver high capacity performance, degrades gracefully during network impairments (such as weather related conditions), and can transport packet or DM traffic in their native formats. Te microwave transport tests functioned as expected, and the test scenarios simulated each of the expected traffic/weather conditions that would experienced in an actual network deployment. Various tests were run to verify that radio link bonding provided the expected 100% increase in the Ethernet (packet) traffic capacity, while maintaining 1+1 subnet circuit protection for the E1 (DM) traffi c flows. Also demonstrated was the capability to support an increased level of best effort packet traffic as well as the bandwidth management functions necessary to reduce the bandwidth allocated to the best effort flow when available bandwidth was reduced – showing that the DM traffic was unaffected and the guaranteed traffic was protected. Since microwave systems are impacted by weather conditions, another important verification was performed that demonstrated the ability of the MINI-LINK, via its adaptive modulation, to show when absolute bandwidth is reduce that only low priority streams are affected. We observed this test in the simulated environment, where the test team increased the attenuation of the signal to reach the modulation switch thresholds, and forced the adaptive modulation functionality to take effect. Te final transport test exercised the operation of the MINI-LINK’s queuing functionality, specifically the “weighted fair queuing”, which offers a more flexible way to schedule on an egress port and prevents star vation of traffic class queues. WFQ can be used together with the Strict Priority scheme to achieve the desire traffi c profiles. Te test was executed using strict priority queuing, which resulted in the loose of certain traffi c flows when an overload condition was reached, in some cases that is the desired result; however, the test was rerun with the system configured with WFQ turn on, and each queue given a percentage of available bandwidth, the results were much different. Instead of dropping data, the queues were services according to their allocations, enabling all traffi c to be serviced. Given the intricacy of this queuing test, please see the detailed test plan for greater details. End user Quality of Experience (QoE)

Tis set of tests show that the end user’s QoE is not severely affected when the MBH and MPBN solutions are confronted with major disturbances, such as overload, link and node failures and other anomolies. Te tests involve technologies such as Ring Protection and queuing as well as VRRP, OSPF, BGP and LDP protocols. One aspect of the QoE demonstration was to demonstrate that the network handles the quality of service and bandwidth requirements of the different use cases in the best possible way when the underlying transmission network is forced to perform under degraded conditions (packet loss, latency and jitter). Multiple network disturbances were injected by the team including: removing a configured link, and monitoring the time required to switch over to a backup path; recover from a fully powered down node; and the insertion of network glitches such as traffic overload to force policing based on traffi c class and priority. Te tests produced the expected results and leveraged the combination of all three of Ericsson’s mobile infrastructure solutions to deliver an end to end solution. Te significant difference between these QoE tests and our previous testing, relates to the combination of all components that comprise the end to end mobile transport network, including the SmartEdge routers, Ethernet switches and the OMS 1410 optical transport systems that provide connectivity and mobile core processing. As with the other scenarios we observed, the network functioned as expected. Te QoE test cases verified that major disturbances in the backbone and transport network comprised of the OMS and MPBN solutions do not create severe impacts on the end-to-end services. It was noted that temporary (induced) degradation of the Quality of Experience was avoided. Te test also shows that the network handles the quality of service and bandwidth requirements of the different use cases properly when the underlying transmission network is forced to operate without certain resources Te backbone and transport network solution complies to these expectations by means of a fully redundant network design with fast fail-over capabilities and a © 2011 Current Analysis Inc. All rights reserved. For more information, please call +1 703 404 9200, toll-free +1 877 787 8947 Europe +33 (0) 1 41 14 83 15. Or visit our Web site: www.currentanalysis.com

6



well designed assignment of quality of service classes and their mapping to priority and weighted queues in the infrastructure equipment. During this section of the demonstration, we observed and in some cases actually caused network faults to demonstrate QoE. We began by creating a link failure, by unplugging an MPLS core link configured between two SmartEdge routers performing the mobile core functions. We noted that the monitoring system identified the failed link which caused the backbone routing process to re-route the traffi c onto other predetermined paths, within 1 second. Te link failure was then reversed, and we observed that the pre-failure routes were reestablished. We were able to note that the test traffi c, which was continuously injected in the baseline profile, was analyzed to determine the time to restore the link was well within the expected sub-second range. Multiple types of link failures we created to demonstrate the networks ability to switchover to alternative paths, restore to pre-failure states and continue to carry traffi c with little or no impact on the end user QoE. Given the resiliency characteristics of the optical network, several tests were conducted to force switchovers within the OMS optical rings. With voice calls active across the network, several link failures were induced. Te switchover time observed were within 50ms; no voice calls were dropped, and the failures did not cause any noticeable glitches in the voice quality of the calls in progress – as experienced via a mobile handset Te eNodeB synch status was observed as remaining in LOCKED mode. Several tests were conducted which required a full shutdown of the OMS, these tests showed minimal disruption to the traffic flows. Perhaps the most impressive QoE test involved demonstrating how the network performed when an entire node failure occurred. Te previous tests dealt with single link failures throughout the network, but this test forced the failure of a complete SmartEdge 1200 backbone router. Te affected traffi c was re-routed onto the other SE1200 in under two seconds. Te average delay noted on the IxN2X test traffic increased after failure due, to additional hop in the traffi c’s path, but the impact on the end-to-end traffi c was negligible. Te process was then reversed, the node was powered on, and after all the routing protocols were restarted, the system reverted to its pre-failure state. Te same scenario was conducted between the redundant SmartEdge 800 nodes service as the interface between the core and the mobile backhaul portion of the network.  Recommended Actio ns

Recommended Vendor Actions

• Ericsson should continue to promote its end to end solutions, which can help mobile operators derived significant benefits in terms of time to market, establishing a flexible and extensible architecture, and reducing the overall operational costs associated with delivering mobile services. Using the end to end message as an umbrella helps operators focus on delivering services over a proven architecture. • Ericsson should also promote the ability of its three mobile solutions to work together in an all Ericsson network as well as the ability for the any one of the three solutions to work independently in a multi-vendor network. Mobile operators seeking to second source their networks will find that they can leverage their solutions from the vendor of choice and still realize the benefits of having a pre-tested network sub-segment (such as the MBH, MPBN or the IP RAN). • Ericsson should provide network operators with all requisite technical documentation and test results it has generated in the normal course of testing their solutions offers. Increased information along these lines could help Ericsson generate additional interest in its solutions from operators early in evaluation and procurement process and help to reduce the time the operators spends in testing the infrastructure. By accepting a baseline set of functionality and performance, operators are then free to focus on service provisioning and aspects of their service delivery that


7



enable them of offer differentiated services. • Ericsson should continue to leverage its professional services capabilities as part of its value to operators. Although the three solutions come verified and pre-tested, operators will seldom have the exact topology and configuration as provided by Ericsson. With the inclusion of professional services, operators can easily customize the network to meet their specific requirements. Further, given a well defined architecture and implementation plan, operators will be capable of incrementally expanding their infrastructures as traffic demand changes and as new services are turned up. • Ericsson should continue to promote its mobile transport infrastructure as meeting the needs of emerging LE services, but stress that it does so not at the expense of supporting legacy services. Equally important is the converse, where an operator needs to support current services but wants to add support for LE services in the future (the future now being 2011 – 2014). Te ability to support multiple generations of mobile traffi c and scale to meet continue growth demands, demonstrates the depth of experience and field proven knowledge the Ericsson brings to the problem. Recommended Buyer Actions

• Mobile operators should evaluate Ericsson as a full service vendor and strategic partner who is capable of providing end to end solutions that are backed up by well know products and services. Operators should press Ericsson for assurances that its products and platforms will continue to be tested and integrated into deliverable solutions as a way to reduce time to market and help them rapidly scale their networks to meet growing bandwidth and service awareness. • Mobile operators should request to receive all requisite technical documentation and test results during the initial evaluation exercises. Tis would enable operators to by-pass much of the unit and system level testing and validation (performed by the vendor), enabling them to focus on providing service differentiation and quality of experience. • Mobile operators looking to upgrade their single generation mobile core network should look closely at Ericsson’s Mobile Packet Backbone Network (MPBN), which provides a transport solution for 2G, 3G and LE mobile core networks. • Mobile operators looking to upgrade their mobile backhaul infrastructures should look closely at Ericsson’s Mobile Backhaul (MBH), which provides the infrastructure to interconnect the MBH and IP Ran. Operators should make sure that their backhaul networks do not become the choke point when providing new high bandwidth services. Operators must also make sure that the backhaul network can deliver the level of QoS and bandwidth management needed to support the uncertainty of rapidly evolving packet based traffi c. • Mobile operators looking to upgrade their mobile transport infrastructures should look closely at Ericsson’s the IP RAN, which provides infrastructure and connectivity of GSM, WCDMA and LE radio networks.

Click the following link for access to the test plan including details regarding the specifics of the test setup such as equipment type, interfaces, and configurations. http://www.currentanalysis.com/t/2011/testspec/


8

E2E Transport Network Solution

Recommend Documents