I studied theoretical aspect of DS-TE and i have learnt that two types of LSP could exist : E-LSP and L-LSP :
For E-LSP, the EXP field of the MPLS shim header conveys both the packet's queue scheduling and drop precedence to each LSR hop, DiffServ is supported by placing packets that require specialized forwarding into different queues. However the bandwidth is reserved for the entire LSP without the PSC-based granularity, and there may be insufficient bandwidth in queues serving some particular PSCs.
Contrary to this, in an L-LSP , the label value is used to infer the queue scheduling for the packet while the drop precedence is carried in the EXP field of the MPLS shim header. Because a label carries the scheduling information, when bandwidth is reserved for a given L-LSP, it is associated with the priority queue to which this LSP belongs.
When I study practice aspect, vendor don't speak about use of E-LSP or L-LSP for DS-TE. According to what I could read here
DS-TE use BC enabling to diffuse available bandwidth of its associated CT (use of sub-pool). This process enable to reserve bandwidth for a particular class of service and guarantee availability on queues serving this class. This process seems to be the application of L-LSP. But during the forwarding phase, the queue scheduling is determined not by MPLS shim header but by exp field.
I don’t understand which sort of LSP is use in practice, and in particular in the case exposed in the cisco document mentioned above.
Someone can help me, please?