[ndnSIM] some doubt about ndnSIM

[Alex Afanasyev]

Hi BoDeng,

NS-3 has its internal mechanism (ns3::Names class) that is used to associate names and pointers, such as pointers to Node objects.  Internally, ns3::Names uses std::map, so it is not exactly hash operations, but should not be too inefficient (each operation on the order of logN).

Creating FIB table from scratch using AddRoute method (this specific version: http://ndnsim.net/doxygen/classns3_1_1ndn_1_1_stack_helper.html#a64db64bd1b98144cda977dc5659a7d98) should be relatively efficient.  There is no way currently to save/restore FIB from the disk, as its data structure is special.

How big is your FIB? 

---
Alex


On Mar 22, 2013, at 1:08 AM, dengbo <dengbo2013 at ict.ac.cn> wrote:

> Dear Alex:
> 
>         My name is BoDeng , I will study for master in September and be student of teacher sun . I have some doubt about your ndnSIM  and 
> wish get your answer.(I am helping teacher sun to do his experiment about ndn with weiwang)
>        The first is about the function Names::Find<Node>
>         (nodeName) used in function  void StackHelper::AddRouter (const std::string &nodeName, 
> const std::string &prefix, uint32_t faceId, int32_t metric). I want to know how you store the nodeName ? Do you use hash?
>      The second is that the fib table is stored in the memory and be added in every experiment , I think it is not suitable when the fib
> table is big , can you store them in disk with some structure so that we can read them from the file and needn`t calculate
> it every time .Also, if the structure is very special , we can create the fib file alone to reduce the experiment time . 
> 
> Thanks
>  -- 
> BSET REGARDS
> DengBo
> Student of Network Technology Research Center , ict , ac
> Addr: No. 6 Kexueyuan South Rd, Haidian, Beijing, China 
> Tel:+86-15271810335

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.lists.cs.ucla.edu/pipermail/ndnsim/attachments/20130325/8b659e29/attachment.html>