<p>In distributed storage systems, node erasures may occur frequently. An <InlineEquation ID="IEq1"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="10623_2025_1697_Article_IEq1.gif" Format="GIF" Height="19" Rendition="HTML" Resolution="72" Type="Linedraw" Width="56" /> </InlineMediaObject> <EquationSource Format="TEX">\((n,k,\ell )\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mo stretchy="false">(</mo> <mi>n</mi> <mo>,</mo> <mi>k</mi> <mo>,</mo> <mi>ℓ</mi> <mo stretchy="false">)</mo> </mrow> </math></EquationSource> </InlineEquation> minimum storage regenerating (MSR) code provides the optimal-bandwidth repair of single node erasure whiling retaining the minimum storage overhead. For multi-node repair, centralized MSR and cooperative MSR codes are proposed to repair <InlineEquation ID="IEq2"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="10623_2025_1697_Article_IEq2.gif" Format="GIF" Height="16" Rendition="HTML" Resolution="72" Type="Linedraw" Width="42" /> </InlineMediaObject> <EquationSource Format="TEX">\(h\ge 2\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi>h</mi> <mo>≥</mo> <mn>2</mn> </mrow> </math></EquationSource> </InlineEquation> node erasures using <i>d</i> helper nodes in a centralized and cooperative way, respectively. <InlineEquation ID="IEq3"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="10623_2025_1697_Article_IEq3.gif" Format="GIF" Height="14" Rendition="HTML" Resolution="72" Type="Linedraw" Width="12" /> </InlineMediaObject> <EquationSource Format="TEX">\(\ell \)</EquationSource> <EquationSource Format="MATHML"><math> <mi>ℓ</mi> </math></EquationSource> </InlineEquation> is called the sub-packetization of the code and <i>d</i> is called repair degree. In a remarkable work, Ye presented cooperative MSR codes (called Ye’s code) with sub-packetization <InlineEquation ID="IEq4"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="10623_2025_1697_Article_IEq4.gif" Format="GIF" Height="19" Rendition="HTML" Resolution="72" Type="Linedraw" Width="74" /> </InlineMediaObject> <EquationSource Format="TEX">\(\mathrm{{exp}}(O(n))\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi mathvariant="normal">exp</mi> <mo stretchy="false">(</mo> <mi>O</mi> <mo stretchy="false">(</mo> <mi>n</mi> <mo stretchy="false">)</mo> <mo stretchy="false">)</mo> </mrow> </math></EquationSource> </InlineEquation> by repeating a previous MSR code for <InlineEquation ID="IEq5"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="10623_2025_1697_Article_IEq5.gif" Format="GIF" Height="15" Rendition="HTML" Resolution="72" Type="Linedraw" Width="74" /> </InlineMediaObject> <EquationSource Format="TEX">\(d-k+h\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi>d</mi> <mo>-</mo> <mi>k</mi> <mo>+</mo> <mi>h</mi> </mrow> </math></EquationSource> </InlineEquation> times. In this paper, motivated by Ye’s code, we find observations to build new centralized multi-node repair schemes of codes with small sub-packetization. The first code is designed for a single pair of (<i>h</i>,&#xa0;<i>d</i>) and the second code is for all possible pairs of (<i>h</i>,&#xa0;<i>d</i>) simultaneously. The main technique adopts a generalized space-sharing method and modifies Ye’s code to get new centralized repair schemes with less space-sharing. The resultant code has sub-packetization no more than <InlineEquation ID="IEq6"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="10623_2025_1697_Article_IEq6.gif" Format="GIF" Height="19" Rendition="HTML" Resolution="72" Type="Linedraw" Width="201" /> </InlineMediaObject> <EquationSource Format="TEX">\(\mathrm{{lcm}}(1,2,\ldots ,n-k)(n-k)^n\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi mathvariant="normal">lcm</mi> <mrow> <mo stretchy="false">(</mo> <mn>1</mn> <mo>,</mo> <mn>2</mn> <mo>,</mo> <mo>…</mo> <mo>,</mo> <mi>n</mi> <mo>-</mo> <mi>k</mi> <mo stretchy="false">)</mo> </mrow> <msup> <mrow> <mo stretchy="false">(</mo> <mi>n</mi> <mo>-</mo> <mi>k</mi> <mo stretchy="false">)</mo> </mrow> <mi>n</mi> </msup> </mrow> </math></EquationSource> </InlineEquation>, which is much smaller than a previous work given by Ye and barg (<InlineEquation ID="IEq7"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="10623_2025_1697_Article_IEq7.gif" Format="GIF" Height="19" Rendition="HTML" Resolution="72" Type="Linedraw" Width="161" /> </InlineMediaObject> <EquationSource Format="TEX">\((\mathrm{{lcm}}(1,2,\ldots ,n-k))^n\)</EquationSource> <EquationSource Format="MATHML"><math> <msup> <mrow> <mo stretchy="false">(</mo> <mi mathvariant="normal">lcm</mi> <mrow> <mo stretchy="false">(</mo> <mn>1</mn> <mo>,</mo> <mn>2</mn> <mo>,</mo> <mo>…</mo> <mo>,</mo> <mi>n</mi> <mo>-</mo> <mi>k</mi> <mo stretchy="false">)</mo> </mrow> <mo stretchy="false">)</mo> </mrow> <mi>n</mi> </msup> </math></EquationSource> </InlineEquation>).</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

New centralized multi-node repair schemes for distributed storage

  • Yaqian Zhang

摘要

In distributed storage systems, node erasures may occur frequently. An \((n,k,\ell )\) ( n , k , ) minimum storage regenerating (MSR) code provides the optimal-bandwidth repair of single node erasure whiling retaining the minimum storage overhead. For multi-node repair, centralized MSR and cooperative MSR codes are proposed to repair \(h\ge 2\) h 2 node erasures using d helper nodes in a centralized and cooperative way, respectively. \(\ell \) is called the sub-packetization of the code and d is called repair degree. In a remarkable work, Ye presented cooperative MSR codes (called Ye’s code) with sub-packetization \(\mathrm{{exp}}(O(n))\) exp ( O ( n ) ) by repeating a previous MSR code for \(d-k+h\) d - k + h times. In this paper, motivated by Ye’s code, we find observations to build new centralized multi-node repair schemes of codes with small sub-packetization. The first code is designed for a single pair of (hd) and the second code is for all possible pairs of (hd) simultaneously. The main technique adopts a generalized space-sharing method and modifies Ye’s code to get new centralized repair schemes with less space-sharing. The resultant code has sub-packetization no more than \(\mathrm{{lcm}}(1,2,\ldots ,n-k)(n-k)^n\) lcm ( 1 , 2 , , n - k ) ( n - k ) n , which is much smaller than a previous work given by Ye and barg ( \((\mathrm{{lcm}}(1,2,\ldots ,n-k))^n\) ( lcm ( 1 , 2 , , n - k ) ) n ).