assign values to a big matrix in R

Question

I have a matrix object, named location, with three columns(ID, latitude, logitude) and 18,289 rows:

#      ID    latitude     longitude  
#  320503    31.29530     120.5735  
#  310104    31.18852     121.4365  
#  310115    31.22152     121.5444  
#  110105    39.92147     116.4431  

I want to calculate the Haversine distance between each ID. In the package geosphere the function distm() can create the symmetric distance matrix, but its dimension is 18289 by 18289. However, R reported the error that it cannot allocate vector of size 2.5 Gb.

Similarly, with the ff package,

data.distance.ff <- ff(0, dim = c(18289, 18289))

produces no errors, but when I assign values to the ff matrix, R produces errors come again.

data.distance.ff[1:18289, 1:18289] <- distm(location[, 2:3] 

Error: cannot allocate vector of size 2.5 Gb
In addition: Warning messages:
1: In matrix(0, ncol = n, nrow = n) :
Reached total allocation of 2047Mb: see help(memory.size)
2: In matrix(0, ncol = n, nrow = n) :
Reached total allocation of 2047Mb: see help(memory.size)
3: In matrix(0, ncol = n, nrow = n) :
Reached total allocation of 2047Mb: see help(memory.size)
4: In matrix(0, ncol = n, nrow = n) :
Reached total allocation of 2047Mb: see help(memory.size)

I can verify this error with:

data.distance.ff[1:10000, 1:10000] <- distm(location[1:10000, 2:3] 

And then get this error:

Error: cannot allocate vector of size 772.1 Mb.

My questions are:

1. Is my code for assigning values to the ff matrix object wrong? Should I be using something special to assign values to an ff object instead?
2. Canff objects handle the storage requirement?
3. Can I use another method to calculate the distance using an apply function that does not involve loops? I know the function distm() produces a matrix twice as large as needed because it is symmetric.
4. Are there any other methods for handling with the big data? The bigmemory package does not appear to work on my Windows computer.