COM.FutureTense.Interfaces
Interface ISyncHash

public interface ISyncHash

This class presents functionality like a hashtable but with cluster sync and LRU behaviors. Note that the use of this object has performance implications. Size should be consider by cost-to-reload data vs memory usage per VM; cluster sync should be used only when a backing store is likely to change and invalidate stored results in a given VM.

Operational considerations (from Netscape) suggest that data read vs write from a synchronized hash should be in a 80/20 ratio if the synchronization is across multiple machines. Non synchronized hashs do not have this suggested restriction. In theory the cost is also low if multiple VM's are on a single machine.

Method Summary

void	clear() Empty a hash - causes cluster sync as needed
java.lang.Object	get(java.lang.String key) Get an object from hash - performs required data validation/VM synchronization checks as needed.
java.lang.String	getName() Get name
java.util.Set	keySet() Return a set of all of the keys in the current hash.
java.lang.Object	put(java.lang.String key, java.lang.Object v) Add something to the hash; will generate appropriate cluster sync methods and LRU behavior.
java.lang.Object	put(java.lang.String key, java.lang.Object v, java.util.Calendar cal) Add something to the hash; will generate appropriate cluster sync methods and LRU behavior.
java.lang.Object	remove(java.lang.String key) Remove an item
int	size() Return the size of the current hash

Method Detail

getName

public java.lang.String getName()

Get name

Returns:

Returns the name of the hash as it is managed. There is no namespace support for hash names.

put

public java.lang.Object put(java.lang.String key,
                            java.lang.Object v)
                     throws java.lang.NullPointerException

Add something to the hash; will generate appropriate cluster sync methods and LRU behavior.

Parameters:

key - Use meaningful, reproducable key names for proper behavior. Limit length; don't use multi-K length keys. 64chars or less is a good guide; 16 chars is optimal for performance but may not be unique.

v - value

Returns:

previous object or null NOTE may toss exception(s) if the create had failed in some unexpected way

Throws:

java.lang.NullPointerException

put

public java.lang.Object put(java.lang.String key,
                            java.lang.Object v,
                            java.util.Calendar cal)

Add something to the hash; will generate appropriate cluster sync methods and LRU behavior.

Parameters:

key - Use meaningful, reproducable key names for proper behavior. Limit length; don't use multi-K length keys. 64chars or less is a good guide; 16 chars is optimal for performance but may not be unique.

v - value

cal - Calendar object specifying the expiration time for this object in the hash. The object added will expire at the time specified by the calendar. Null is not allowed.

Returns:

previous object or null NOTE may toss exception(s) if the create had failed in some unexpected way

Since:

6.1

get

public java.lang.Object get(java.lang.String key)

Get an object from hash - performs required data validation/VM synchronization checks as needed.

Parameters:

key -

clear

public void clear()

Empty a hash - causes cluster sync as needed

remove

public java.lang.Object remove(java.lang.String key)

Remove an item

Parameters:

key -

Returns:

the object, null if not found. Causes cluster sync as needed. It is more efficient to clear() a hash then to remove all entries one by one; first is loop cost, second is cluster-sync cost which can be very high.

size

public int size()

Return the size of the current hash

Returns:

int size

Since:

6.1

keySet

public java.util.Set keySet()

Return a set of all of the keys in the current hash.

Returns:

Set of string keys; never null

Since:

6.1