IMS and ATCA James Rafferty James raffertydialogic com

  • Slides: 28
Download presentation

IMS and ATCA James Rafferty James. rafferty@dialogic. com

IMS and ATCA James Rafferty James. rafferty@dialogic. com

3 Agenda • • • Open Systems for Service Providers SIP: A Success Story

3 Agenda • • • Open Systems for Service Providers SIP: A Success Story IP Multimedia Subsystem ATCA Where are we on the adoption curve? IMS and ATCA: The Road Ahead

4 Open Systems for Service Providers • Traditional switches (pre-IP) built on proprietary hardware

4 Open Systems for Service Providers • Traditional switches (pre-IP) built on proprietary hardware and various quasi-standard telecom protocols • With advent of IP, a chance to revisit the model • Two key directions emerged: – Open protocols – Open hardware • How is that working out?

5 Open Protocols • Good News – Open protocols are very popular – But,

5 Open Protocols • Good News – Open protocols are very popular – But, there’s been too many of them • Circuit-Based: MFR 2, ISDN PRI, SS 7 • H. 323, MGCP, H. 248, SIP – From a variety of standards groups and consortia • ITU-T, ANSI, ETSI, IETF, 3 GPP, W 3 C, … – Fortunately…

6 Winner of the Standards Wars • SIP (Session Initiation Protocol) • Two Key

6 Winner of the Standards Wars • SIP (Session Initiation Protocol) • Two Key Endorsements – Microsoft – Third Generation Partnership Program (3 GPP)

7 SIP and Open Telecom • Since 2001, Vo. IP industry has built products

7 SIP and Open Telecom • Since 2001, Vo. IP industry has built products that are SIP enabled • But SIP is pretty complex – Many core standards – Many standard extensions – Many proprietary “de facto” extensions • How to sort out this puzzle?

8 One Answer: IMS • Wireless Vendors worked to produce standards for Multimedia Services

8 One Answer: IMS • Wireless Vendors worked to produce standards for Multimedia Services over SIP • Design Criteria: – – Provide common architecture for multiple services IP at the core Offer access to users on existing networks Need to interwork between IP and circuit switched at the edges • Result: IP Multimedia Subsystem (IMS) – Developed by Third Generation Partnership Project (3 GPP) – Endorsed by both wireless and wireline industry groups

IMS Architecture

IMS Architecture

IP Multimedia Subsystem (IMS) Simplified View Key Elements: Application Servers § AS – Application

IP Multimedia Subsystem (IMS) Simplified View Key Elements: Application Servers § AS – Application Server § SCIM - Service Capability Interaction Manager SCIM § MRFC - Multimedia Resource Function Controller § MRFP - Multimedia Resource Function Processor HSS/HLR S-CSCF SIP C I-CSCF SCFF § CSCF- Call Session Control Function SIP § BGCF - Breakout Gateway Control Function BGCF § MGCF - Media Gateway Control Function P-CSCF MRFC CSCF § MRF – Media Resource Function § MGW - Media Gateway MGCF § HSS - Home Subscription Server § HLR - Home Location Register MRFP MRF RTP MGW SS 7

11 Benefits of IMS • Reference Architecture for Vendors and Customers • Provides standard

11 Benefits of IMS • Reference Architecture for Vendors and Customers • Provides standard profile of SIP – 3 GPP standards reference core SIP standards and extensions • No work in process documents are mandated • Provides focus for SIP and related interop • Updated about every year and one half – Updates reflect progress in filling standards gaps

IMS: Foundation for Multimedia Services Unified Messaging Announcements Pre-Paid 3 G Voice Mail SIP

IMS: Foundation for Multimedia Services Unified Messaging Announcements Pre-Paid 3 G Voice Mail SIP S-CSCF/ SCIM IP MSC PSTN RTP MGCF/ MGW Video Conferencing IP Call Center Conferencing Video Ringback Gaming SIP MSCML SIP 2. 5 G Video Mail SIP Routing Cloud SIP SIP SIP Voice. XML Quick. Time w/ Voice. XML HTTP w/ MSCML Web Content NFS IP Phones MRF Text MPEG-4 FTP CMTS Cable mp 3 Network Storage

13 IMS: Filling out the Specs • Work continues on notable IMS components •

13 IMS: Filling out the Specs • Work continues on notable IMS components • Example: Internet Engineering Task Force is developing a SIP based media control protocol – Likely to be adopted by 3 GPP for IMS – Will provide a linchpin in the critical interaction between Application Servers and Media Servers • Current IMS spec is Release 7 (2007) • Next IMS Release 8 is targeted for 2008

14 IMS and Sorting Out SIP • The promise of IMS is a reference

14 IMS and Sorting Out SIP • The promise of IMS is a reference architecture that many vendors buy into • Reality is more complex • Vendors ARE deploying pre-IMS or partial IMS solutions – SIP is widely used, but interoperability issues remain – IMS will help encourage standard profiles of SIP – As a vendor, we’re already seeing more pressure to adapt the SIP specs endorsed by 3 GPP

15 What about Open Hardware? • PICMG devised the first pass of a telecom

15 What about Open Hardware? • PICMG devised the first pass of a telecom architecture early this century – Used Compact PCI – Had Ethernet backbone – PICMG 2. 16 – Enabled use of bladed architecture with IP protocols tying together the blades • Reaction of Telecom Vendors – Not sufficiently carrier grade – PCI bus is single point of failure

16 ATCA to the Rescue • PICMG started all over again • Created the

16 ATCA to the Rescue • PICMG started all over again • Created the Advanced Telecommunications Computing Architecture (ATCA) • Goals: – – – Support very high bandwidth backbones in a chassis Provide specifications for all key hardware components Support multiple switch fabrics Include built-in management tools Build in robust high availability hardware from the start

30 -Nov-20 ATCA Expanded Board Size 8 U Advanced. TCA 125 square inches ~200

30 -Nov-20 ATCA Expanded Board Size 8 U Advanced. TCA 125 square inches ~200 watts power 6 U 160 mm Connector Zones 55 square inches ~50 watts power Connector Zones Compact. PCI 280 mm

30 -Nov-20 Bus vs. Switched Fabric Comparison Shared Bus can jam with one failure

30 -Nov-20 Bus vs. Switched Fabric Comparison Shared Bus can jam with one failure Switched Fabric can have redundant paths

Example: Evolution to ATCA Compact PCI Boards ATCA Carrier Card Add HW Assist Rackmount

Example: Evolution to ATCA Compact PCI Boards ATCA Carrier Card Add HW Assist Rackmount Server ATCA Host Blade Software-Based Media Server ATCA AMC Cards ATCA Chassis

20 ATCA: Vision and Reality • ATCA Vision of Open Hardware is a great

20 ATCA: Vision and Reality • ATCA Vision of Open Hardware is a great idea – Encourage creation of open hardware eco-system – Get broad participation of experts in all key components – Build best of breed systems using open components • The Reality? – The x. TCA eco-system has partially taken hold – Lots of competition from private bladed architectures – All of the fabric choices fragmented the vendor implementations – Many vendors and carriers have not bought in

21 In Perspective • Can you buy a deployable IMS implementation using x. TCA

21 In Perspective • Can you buy a deployable IMS implementation using x. TCA components today? • In total: – No • In parts: – Yes • Does this mean that Open Systems are a failure? – No, not at all

22 Where are we on the adoption curve? • For SIP – Core SIP

22 Where are we on the adoption curve? • For SIP – Core SIP RFCs widely adopted – Lots of proven interop between components – Interop via SIPIT and other industry assocations • For IMS – IMS concepts being used, but standards are incomplete – Hence, adoption of pre-IMS built around SIP • For ATCA – Sporadic adoption by key vendors – Alternatives have emerged

23 Alternatives to ATCA • Blade Servers – Platforms developed by several computer vendors

23 Alternatives to ATCA • Blade Servers – Platforms developed by several computer vendors – Heavy processing focus; less emphasis on I/O throughput • Rack Mounted Servers – Use server as appliance approach – 1 and 2 U form factors; expand via rack and stack – Purpose built hardware coupled with open protocols

24 Open Systems Example • IP Media Servers have been adopted in several form

24 Open Systems Example • IP Media Servers have been adopted in several form factors – Began as proprietary, carrier rack systems – Shift toward 1 and 2 U appliances, often with standard CPUs and no DSPs – Blade servers offer high density and scalability – ATCA processor blades can also run media server code and scale up via addition of more blades – Micro. TCA will enable smaller form factor media servers built on x. TCA eco-system

25 So What about ATCA and IMS? • ATCA has helped spawn an open

25 So What about ATCA and IMS? • ATCA has helped spawn an open hardware movement – But rack mount servers and blade servers have been hugely popular • IMS is one of the factors driving wide adoption of SIP – Open protocols have been a big winner, with SIP being the focus – IMS offers a reference architecture for SIP and ways of gluing SIP components together

26 ATCA and IMS: Evolution, not Revolution • ATCA still has an uphill climb

26 ATCA and IMS: Evolution, not Revolution • ATCA still has an uphill climb • Service providers want reliable HA architectures • ATCA has the promise, but needs more adoption – Smaller x. TCA form factors may be more successful – Can build on same eco-system as ATCA • IMS has lots of vendor mind share, but… – Specifications are still maturing – Still early on the adoption curve – Vendors are using IMS concepts

27 Resources • IMS Forum – www. imsforum. org • 3 GPP – www.

27 Resources • IMS Forum – www. imsforum. org • 3 GPP – www. 3 GPP. org • Internet Engineering Task Force – www. ietf. org • ETSI TISPAN http: //portal. etsi. org/portal_common/home. asp? tbkey 1=TISPAN • PICMG – www. picmg. org • Multi. Service Forum – www. msforum. org

28 Summary • Movement toward Open Standards is strong in Telecom • SIP has

28 Summary • Movement toward Open Standards is strong in Telecom • SIP has been widely adopted • IMS offers promise as a reference architecture, but still in early stages of deployment • ATCA has supporters but is vying with alternatives as the basis for next generation hardware • Combination of ATCA and IMS makes sense, but is not happening yet on a wide scale