Virtual Network Embedding: A Hybrid Vertex Mapping Solution for Dynamic Resource Allocation

Table 2

Solution to VNE constraints and approaches adopted by various proposals.

Reference	ABC		VN requests	OM	Adaptability	Method		Opt obj
	C	NC				Vertex	Edge
[9]	✓		ONL	Cent	PM	GNM	SPM/MCF	MRU
[14]	✓		ONL	Cent	LF	IM	SPM	MQR
[15]	✓		ONL	Dist	NLF, CUR	IM	SPM	MAT
[16]		NCap, AC	OFL	Cent	NC	IM	SPM	MNC
[17]		NCap, AC	ONL	Cent	CUR	SA	SPM	MOUC
[18]		AC	OFL	Cent	CNC	GNM	SPM	MRU
[13]	✓		ONL	Cent	NC	D-Vine/R-Vine	MCF	MARR
[19]		NCap	OFL	Cent	NC	—	MCF	MAR
[20]		AC	OFL	Dist	NC	IM	SPM	MNC
[21]	✓		ONL	Cent	NC	SA	SPM	MMT
[10]	✓		ONL	Cent	Re-opt	GNM/ D-Vine/R-Vine	SPM/MCF	MAR

Where in Table 2—ABC: all basic constraints data, NC: not considered, AC: admission control, OFL: offline, Dist: distributed, LF: link failures, CUR: change in user’s requirements, IM: iterative method, GNM: greedy node mapping, D-Vine: deterministic rounding based virtual network embedding algorithm, R-Vine: randomized rounding based virtual network embedding algorithm, MCF: multicommodity flow, Opt Obj: optimization objective, MQR: maximize quality and resilience, MNC: minimize network cost, MARR: maximize acceptance ratio and revenue, MAR: maximize acceptance ratio, C: considered, NCap: node capacity, ONL: online, Cent: centralized, PM: path migration, NLF: node and link failures, CNC: change in network conditions, SA: simulated annealing, SPM: shortest path mapping, TOM: type of mapping, MRU: maximize resource usage, MAT: minimize adaptation time, MOUC: minimize overlay usage cost, Re-opt: reoptimization, MMT: minimize mapping time.