<p>In the present paper, we introduce a message-recovery attack based on the Modular Knapsack Problem, applicable to all variants of the NTRU-HPS cryptosystem. Assuming that a fraction <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(\epsilon \)</EquationSource> <EquationSource Format="MATHML"><math> <mi>ϵ</mi> </math></EquationSource> </InlineEquation> of the coefficients of the message <i>m</i><InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(\in \{-1,0,1\}^N\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mo>∈</mo> <msup> <mrow> <mo stretchy="false">{</mo> <mo>-</mo> <mn>1</mn> <mo>,</mo> <mn>0</mn> <mo>,</mo> <mn>1</mn> <mo stretchy="false">}</mo> </mrow> <mi>N</mi> </msup> </mrow> </math></EquationSource> </InlineEquation> and of the nonce vector <InlineEquation ID="IEq3"> <EquationSource Format="TEX">\(\textbf{r} \in \)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi mathvariant="bold">r</mi> <mo>∈</mo> </mrow> </math></EquationSource> </InlineEquation> <InlineEquation ID="IEq4"> <EquationSource Format="TEX">\(\{-1,0,1\}^N\)</EquationSource> <EquationSource Format="MATHML"><math> <msup> <mrow> <mo stretchy="false">{</mo> <mo>-</mo> <mn>1</mn> <mo>,</mo> <mn>0</mn> <mo>,</mo> <mn>1</mn> <mo stretchy="false">}</mo> </mrow> <mi>N</mi> </msup> </math></EquationSource> </InlineEquation> are known in advance at random positions, we reduce message decryption to finding a short vector in a lattice that encodes an instance of a modular knapsack system. This allows us to address a key question: how much information about <InlineEquation ID="IEq5"> <EquationSource Format="TEX">\(\textbf{m},\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi mathvariant="bold">m</mi> <mo>,</mo> </mrow> </math></EquationSource> </InlineEquation> or about the pair <InlineEquation ID="IEq6"> <EquationSource Format="TEX">\((\textbf{m},\textbf{r}),\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mo stretchy="false">(</mo> <mi mathvariant="bold">m</mi> <mo>,</mo> <mi mathvariant="bold">r</mi> <mo stretchy="false">)</mo> <mo>,</mo> </mrow> </math></EquationSource> </InlineEquation> is required before recovery becomes feasible? A FLATTER reduction successfully recovers the message, in practice when <InlineEquation ID="IEq7"> <EquationSource Format="TEX">\(\epsilon \approx 0.45.\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi>ϵ</mi> <mo>≈</mo> <mn>0.45</mn> <mo>.</mo> </mrow> </math></EquationSource> </InlineEquation> Our implementation finds <InlineEquation ID="IEq8"> <EquationSource Format="TEX">\(\textbf{m}\)</EquationSource> <EquationSource Format="MATHML"><math> <mi mathvariant="bold">m</mi> </math></EquationSource> </InlineEquation> within a few minutes on a commodity desktop.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Message recovery attack in NTRU via knapsack

  • Eirini D. Poimenidou,
  • Konstantinos A. Draziotis

摘要

In the present paper, we introduce a message-recovery attack based on the Modular Knapsack Problem, applicable to all variants of the NTRU-HPS cryptosystem. Assuming that a fraction \(\epsilon \) ϵ of the coefficients of the message m \(\in \{-1,0,1\}^N\) { - 1 , 0 , 1 } N and of the nonce vector \(\textbf{r} \in \) r \(\{-1,0,1\}^N\) { - 1 , 0 , 1 } N are known in advance at random positions, we reduce message decryption to finding a short vector in a lattice that encodes an instance of a modular knapsack system. This allows us to address a key question: how much information about \(\textbf{m},\) m , or about the pair \((\textbf{m},\textbf{r}),\) ( m , r ) , is required before recovery becomes feasible? A FLATTER reduction successfully recovers the message, in practice when \(\epsilon \approx 0.45.\) ϵ 0.45 . Our implementation finds \(\textbf{m}\) m within a few minutes on a commodity desktop.