Recent events have highlighted the need to balance security with dataset privacy. To improve this... more Recent events have highlighted the need to balance security with dataset privacy. To improve this balance, a data owner and data querier should be able to clearly articulate what will be shared and ensure that only this information is shared. This guarantee should hold even if one of the parties misbehaves. More precisely, the parties should be able to agree on a policy for what type of questions should be asked. Then the data querier should only learn results of allowed queries and no information about irrelevant data. The data owner should be assured that the policy is properly enforced, but learn nothing about individual queries. Proper implementation of controlled and private information sharing has multiple use cases in the U.S. government. Current approaches compromise the privacy needs of either the data owner or data querier. Cryptographic approaches can solve this problem in theory. Tools such as multi-party computation, homomorphic encryption, and symmetric searchable encryption allow parties to participate in some form of controlled information sharing. These tools have significant practicality and functionality limitations. The Intelligence Advanced Research Projects Activity created the Security and Privacy Assurance Research program or SPAR to develop practical approaches to controlled data sharing. 1 This document describes the current state of SPAR Technical Area 1: practical and private database access. The goal is to allow the data querier to query privately a sensitive database managed by the data owner. These technologies had a very aggressive performance goal: no more than five times worse than an unprotected system. Achieving this vision would be transformative, enabling a wide variety of information sharing applications. SPAR technology has been developed and evaluated in two settings: a research program and a pilot demonstration. The original research program developed multiple technologies with very different approaches. These technologies support various query types common in database applications. The evaluation contained two major components, evaluating: 1) the cryptographic security of the underlying technology 2) the performance of the individual query types on a software prototype. This evaluation used synthetic data and focused on performance of individual query types. Teams largely succeeded in meeting the aggressive performance and security requirements. Multiple technologies with strong security had query response times within three times of an unprotected system for at a 10 terabyte scale for important query types, faster than the goal set in the program vision. The original research program could not predict how SPAR technology would perform in a real use case. There are crucial aspects of any use case that are difficult to predict: 1) operators' opinions on the tradeoff between privacy and performance 2) real data's effect on performance 3) how the systems would respond to realistic queries 4) what policy controls need to be enforced. To answer these questions for a representative and important use case, SPAR technology was demonstrated in a pilot U.S. Government use case in July 2015. 2 U.S. Government opera-
You will not find it difficult to prove that battles, campaigns, and even wars have been won or l... more You will not find it difficult to prove that battles, campaigns, and even wars have been won or lost primarily because of logistics. History is rich with examples that showcase the power of military logistics and its influence in the outcome of wars. Hannibal crossing the Alps with foot soldiers, horsemen, and elephants to gain a string of victories in central Italy between 218 and 204 BCE relied on logis-tical planning, cutting supply lines for Roman forces and seizing Roman supply depots [1]. The six years of the Battle of the Atlantic in World War II were a struggle to get a million tons of imported material to Britain every week, fighting German efforts to sink as many of the cargo ships as possible [2]. As the battle raged in the Atlantic Ocean, Allied bombers were destroying German access to oil refineries and synthetic fuel factories. By 1944, the Germans did not have enough fuel for aircraft to protect the oil facilities that remained or for the fleet of submarines that had ...
Several DES-like ciphers aren't utilizing their full potential strength, because of the short key... more Several DES-like ciphers aren't utilizing their full potential strength, because of the short key and linear or otherwise easily tractable algorithms they use to generate their key schedules. Using DES as example, we show a way to generate round subkeys to increase the cipher strength substantially by making relations between the round subkeys practically intractable. £ 1. Given any ¥ bits of the ¦ round keys, derived from an unknown generating key (where ¥ is less than the total amount of round keys material), it is hard to find any of the remaining key bits from the ¥ known bits.
Several DES-like ciphers aren't utilizing their full potential strength, because of the short... more Several DES-like ciphers aren't utilizing their full potential strength, because of the short key and linear or otherwise easily tractable algorithms they use to generate their key schedules. Using DES as example, we show a way to generate round subkeys to increase the cipher strength substantially by making relations between the round subkeys practically intractable.
Recent events have highlighted the need to balance security with dataset privacy. To improve this... more Recent events have highlighted the need to balance security with dataset privacy. To improve this balance, a data owner and data querier should be able to clearly articulate what will be shared and ensure that only this information is shared. This guarantee should hold even if one of the parties misbehaves. More precisely, the parties should be able to agree on a policy for what type of questions should be asked. Then the data querier should only learn results of allowed queries and no information about irrelevant data. The data owner should be assured that the policy is properly enforced, but learn nothing about individual queries. Proper implementation of controlled and private information sharing has multiple use cases in the U.S. government. Current approaches compromise the privacy needs of either the data owner or data querier. Cryptographic approaches can solve this problem in theory. Tools such as multi-party computation, homomorphic encryption, and symmetric searchable encryption allow parties to participate in some form of controlled information sharing. These tools have significant practicality and functionality limitations. The Intelligence Advanced Research Projects Activity created the Security and Privacy Assurance Research program or SPAR to develop practical approaches to controlled data sharing. 1 This document describes the current state of SPAR Technical Area 1: practical and private database access. The goal is to allow the data querier to query privately a sensitive database managed by the data owner. These technologies had a very aggressive performance goal: no more than five times worse than an unprotected system. Achieving this vision would be transformative, enabling a wide variety of information sharing applications. SPAR technology has been developed and evaluated in two settings: a research program and a pilot demonstration. The original research program developed multiple technologies with very different approaches. These technologies support various query types common in database applications. The evaluation contained two major components, evaluating: 1) the cryptographic security of the underlying technology 2) the performance of the individual query types on a software prototype. This evaluation used synthetic data and focused on performance of individual query types. Teams largely succeeded in meeting the aggressive performance and security requirements. Multiple technologies with strong security had query response times within three times of an unprotected system for at a 10 terabyte scale for important query types, faster than the goal set in the program vision. The original research program could not predict how SPAR technology would perform in a real use case. There are crucial aspects of any use case that are difficult to predict: 1) operators' opinions on the tradeoff between privacy and performance 2) real data's effect on performance 3) how the systems would respond to realistic queries 4) what policy controls need to be enforced. To answer these questions for a representative and important use case, SPAR technology was demonstrated in a pilot U.S. Government use case in July 2015. 2 U.S. Government opera-
You will not find it difficult to prove that battles, campaigns, and even wars have been won or l... more You will not find it difficult to prove that battles, campaigns, and even wars have been won or lost primarily because of logistics. History is rich with examples that showcase the power of military logistics and its influence in the outcome of wars. Hannibal crossing the Alps with foot soldiers, horsemen, and elephants to gain a string of victories in central Italy between 218 and 204 BCE relied on logis-tical planning, cutting supply lines for Roman forces and seizing Roman supply depots [1]. The six years of the Battle of the Atlantic in World War II were a struggle to get a million tons of imported material to Britain every week, fighting German efforts to sink as many of the cargo ships as possible [2]. As the battle raged in the Atlantic Ocean, Allied bombers were destroying German access to oil refineries and synthetic fuel factories. By 1944, the Germans did not have enough fuel for aircraft to protect the oil facilities that remained or for the fleet of submarines that had ...
Several DES-like ciphers aren't utilizing their full potential strength, because of the short key... more Several DES-like ciphers aren't utilizing their full potential strength, because of the short key and linear or otherwise easily tractable algorithms they use to generate their key schedules. Using DES as example, we show a way to generate round subkeys to increase the cipher strength substantially by making relations between the round subkeys practically intractable. £ 1. Given any ¥ bits of the ¦ round keys, derived from an unknown generating key (where ¥ is less than the total amount of round keys material), it is hard to find any of the remaining key bits from the ¥ known bits.
Several DES-like ciphers aren't utilizing their full potential strength, because of the short... more Several DES-like ciphers aren't utilizing their full potential strength, because of the short key and linear or otherwise easily tractable algorithms they use to generate their key schedules. Using DES as example, we show a way to generate round subkeys to increase the cipher strength substantially by making relations between the round subkeys practically intractable.
Uploads
Papers by Uri Blumenthal