VPN optimization by defragmentation and deduplication apparatus and method
View Patent ↗An apparatus for optimizing a virtual private network operates by defragmenting and deduplicating transfer of variable sized blocks. A large data object is converted to a plurality of data paragraphs by a fingerprinting method. Each data paragraph is cached and hashed. The hashes are transmitted between a primary and a satellite apparatus. Only data paragraphs which are not cached at both the primary and satellite are transferred. The data object is integrated from data paragraphs stored in cache and transmitted to its destination IP address.
1. A method for operating a system to optimize a virtual private network by defragmentation and deduplication of large data transmissions, the system comprising: a wide area network coupling a satellite apparatus to a primary apparatus, a satellite apparatus comprising: a circuit to receive a data object and determine a plurality of data paragraphs, a circuit to compute a hash for each data paragraph, a circuit to pack a packet with a plurality of hashes; a primary apparatus comprising: a data paragraph cache store, a circuit to receive a packed packet from a satellite apparatus, a circuit to integrate data paragraphs into a data object, and network interfaces and means for sending and receiving TCP packets in each apparatus; the method comprising:
opening a virtual private network connection between a satellite apparatus and a primary apparatus,
establishing a persistent TCP connection between a satellite apparatus and a primary apparatus through a wide area network,
setting a maximum TCP buffer size at both the satellite apparatus and the primary apparatus,
setting a TCP window scale to support large transmission sizes,
transmitting at least one packed packet comprising a plurality of hashes and a destination IP address between a satellite apparatus and a primary apparatus,
within the primary apparatus:
receiving data paragraphs and hashes from a satellite apparatus until all data paragraphs are available,
receiving a packed packet comprising a plurality of hashes and a destination IP address,
integrating a data object by retrieving data paragraphs from data paragraph cache store and
transmitting the data object to the destination IP address;
within the satellite apparatus:
receiving a data object and a destination IP address,
determining a plurality of data paragraphs by computing fingerprints on the data object,
packing a packet with hashes computed from each data paragraph and with the destination IP address,
checking a transfer diary to determine if data paragraphs are available in the primary cache,
queuing data paragraphs for transmission to primary cache,
transmitting a hash to a primary apparatus for each data paragraph queued for transmission to said primary apparatus,
determining that all data paragraphs required for a data object are available at the primary apparatus, and
transmitting the packed packet to the primary apparatus.
2. The method of claim 1 wherein determining a plurality of data paragraphs by computing fingerprints on the data object comprises:
within the satellite apparatus:
incrementally reading a data object as a string of text,
delimiting a data paragraph after a minimum length,
delimiting a data paragraph at a maximum length,
delimiting a data paragraph upon matching a fingerprint to a masked key prior to a maximum length,
computing the fingerprint upon reading each additional byte of the data object after a minimum length and prior to a maximum length, and
delimiting a data paragraph upon reaching the end of the data object.
3. The method of claim 1 wherein receiving data paragraphs and hashes from a satellite apparatus until all data paragraphs are available, comprises
within the primary apparatus:
receiving a hash from a satellite apparatus,
checking a diary store to determine if a data paragraph having a hash is available,
receiving and storing data paragraphs into data paragraph cache store if needed, and
signaling to a satellite apparatus when all queued data paragraphs are available in data paragraph cache store.
4. The method of claim 1 further comprising:
receiving a plurality of packets having a common destination IP corresponding to fragmentation of a data object intended to be transmitted through a virtual private network from a source in a local area network coupled to the satellite apparatus to a destination IP address in a local area network coupled to the primary apparatus,
terminating the connection, and
defragmenting the data object.
5. The method of claim 1 further comprising:
determining that a data paragraph queued for transmission from a satellite apparatus to a primary apparatus is unnecessary and
signaling the satellite apparatus to remove it from the transmission queue.