Slide title In CAPITALS 50 pt Slide subtitle 32 pt Issues with CMIPv6 Suresh Krishnan
Top right corner for field-mark, customer or partner logotypes. See Best practice for example. Slide title 40 pt Slide subtitle 24 pt Text 24 pt Bullets level pt Ericsson Confidential Barriers to deployment CMIPv6 has not been widely deployed –Even though a mature standard has existed for a long time This is due to several factors –Implementation issues (I) –Deployment issues (D) –Operational issues (O) –Layer 9 issues (Human/Political/FUD/…) (9)
Top right corner for field-mark, customer or partner logotypes. See Best practice for example. Slide title 40 pt Slide subtitle 24 pt Text 24 pt Bullets level pt Ericsson Confidential Delay in deployment of IPv6 (D) IPv6 has been standardized more than a decade ago but there has been negligible deployment of IPv6 –This has started to change recently CMIPv6 requires IPv6 to be deployed Hence, CMIPv6 deployments have also been delayed –DSMIPv6 (RFC5555) will probably help resolve this dependency Allows deployment over IPv4 networks
Top right corner for field-mark, customer or partner logotypes. See Best practice for example. Slide title 40 pt Slide subtitle 24 pt Text 24 pt Bullets level pt Ericsson Confidential Non-inclusion in SDO Architectures (9) There are several SDOs defining network architectures for mobile networks –3GPP,3GPP2, Wimax Forum… Until recently CMIPv6 has not been chosen as a mobility solution by any other SDO –They either elected to create their own protocols (eg. GTP) (or) –They picked other IETF mobility protocols (MIPv4,PMIP etc.) Hence no pressure on vendors to implement CMIPv6 on their equipment
Top right corner for field-mark, customer or partner logotypes. See Best practice for example. Slide title 40 pt Slide subtitle 24 pt Text 24 pt Bullets level pt Ericsson Confidential Complex security mechanisms (I,D) The original CMIPv6 spec mandates the use of IPsec and IKEv1 for securing signaling between the MN and the HA (since updated to use IKEv2) –The implementation of the CMIPv6 to IPsec interface is very difficult since it is barely specified This leads to high implementation cost and complexity –The MN-HA security model requires the setup of one SA per MN This makes it very hard to deploy
Top right corner for field-mark, customer or partner logotypes. See Best practice for example. Slide title 40 pt Slide subtitle 24 pt Text 24 pt Bullets level pt Ericsson Confidential Lack of coherent architecture (I,D,O,9) There have been several extensions to CMIPv6 –FMIPv6 for fast handovers –HMIPv6 for hierarchical mobility mgmt –OMIPv6 for enhanced RO –… But, there is no cohesive mobility architecture that holds these extensions together –There is no exhaustive catalog of these extensions –It is unclear whether these extensions would work together or not (or even if they intended to work together) This causes a lot of confusion and incompatible implementations
Top right corner for field-mark, customer or partner logotypes. See Best practice for example. Slide title 40 pt Slide subtitle 24 pt Text 24 pt Bullets level pt Ericsson Confidential Signaling and tunneling overhead (D,O,9) CMIPv6 requires signaling messages to be sent from the MN after each handover –It also requires periodic refresh messages To refresh bindings at the HA To refresh key material for route optimized traffic signaling Every data packet sent by the MN consumes an additional IPv6 header leading to extra overhead –40 bytes extra constitutes a large overhead (percentage wise) for small packets like VOIP This overhead is typically incurred on a radio link where the spectrum is both limited and extremely expensive
Top right corner for field-mark, customer or partner logotypes. See Best practice for example. Slide title 40 pt Slide subtitle 24 pt Text 24 pt Bullets level pt Ericsson Confidential Lack of network participation (O,9) Operators desire to be in control of the mobility management –Since CMIPv6 does not involve entities in the access network, operators perceive this to be a threat –May desire to make a multi-access terminal connect to a lower bit-cost link or a less congested link when available Network node may have additional information about the dynamic conditions –This information can be utilized by network-assisted mobility protocols or sent to the MN as an update of an existing routing rule to make it reflect the dynamic conditions of the network
Top right corner for field-mark, customer or partner logotypes. See Best practice for example. Slide title 40 pt Slide subtitle 24 pt Text 24 pt Bullets level pt Ericsson Confidential Network initiated mobility (PMIPv6) PMIPv6 was touted as the perfect solution for these problems But unfortunately, it came with its own bag of issues –Poor (almost non-existent) multi-access support –Issues with inter-technology handovers –No support for simultaneous multi-access (SMA) –Higher opertational complexity in the network
Top right corner for field-mark, customer or partner logotypes. See Best practice for example. Slide title 40 pt Slide subtitle 24 pt Text 24 pt Bullets level pt Ericsson Confidential Way forward We need a new mobilty solution that draws upon the strengths of CMIPv6 and the strengths of PMIPv6 We need a coherent architecture document that ties several extensions together
Top right corner for field-mark, customer or partner logotypes. See Best practice for example. Slide title 40 pt Slide subtitle 24 pt Text 24 pt Bullets level pt Ericsson Confidential