Many works [16-18] have discussed the kNN and CkNN queries on the distributed and mobile environments like mobile ad hoc NETworks (MANETs) and wireless sensors networks (WSN).
 proposed a Voronoi-based approach for CKNN query.
The continuous k nearest neighbors search (CkNN Search)  is a variation of kNN query.
Using Voronoi diagram (VD) for CkNN search is proposed in the pull-based approach .
Hence, one of the objectives of the designed algorithm, DCkNN, is to minimize the update cost for the CkNN query.
In the mobile and distributed environments, like WSNs or MANETS, the pull-based approach may not be a good match for CkNN search due to the problems discussed in Section 2.
To the best of our knowledge,  is the only approach that uses network distances to find CKNN.
Next we list our observations on the cache design for CKNN or RCKNN problem.
* CKNN or RCKNN: The storage is similar to KNNs except that instead of a single point, the entire line segment between which the CKNNs or RCKNNs lie, is stored.
In the second experiment, we studied the effect of increasing the value of K and the densities of the interest points on the total numbers of ESPs (Element Split Points) and OSPs (Object Split Points.) We present the average results of 20 runs of enhanced continuous K nearest neighbor queries (CKNN) where K varied from 2 to 10 for different entities (hospitals, restaurants, ...
CKNNs or RCKNNs for a particular segment under consideration, which are also a finite set of nodes.
The proposed metric LMNN + DML using CkNN
is compared with the existing metric learning algorithms LMNN  and LDML  with the classification mechanism MkNN.