安恒信息連續(xù)被Gartner®列為“數(shù)據(jù)分類分級(jí)領(lǐng)域”代表廠商
?Gartner
此前,在《2022中國網(wǎng)絡(luò)安全技術(shù)成熟度曲線(Hype Cycle for Security in China, 2022)》報(bào)告中,安恒信息被列為數(shù)據(jù)安全平臺(tái)、云安全資源池、智慧城市網(wǎng)絡(luò)物理系統(tǒng)安全、態(tài)勢感知、數(shù)據(jù)分類分級(jí)和攻防團(tuán)隊(duì)六大領(lǐng)域代表廠商。
在《2023中國網(wǎng)絡(luò)安全技術(shù)成熟度曲線(Hype Cycle for Security in China, 2023)》報(bào)告中,安恒信息被列為數(shù)據(jù)安全平臺(tái)、CPS安全、云安全資源庫、軟件成分分析、數(shù)據(jù)分類、攻防對(duì)抗六大領(lǐng)域標(biāo)桿供應(yīng)商。
伴隨數(shù)字中國建設(shè)的持續(xù)深入,海量數(shù)據(jù)呈幾何級(jí)數(shù)增長,數(shù)據(jù)應(yīng)用場景持續(xù)拓展,數(shù)據(jù)分類分級(jí)的重要性愈發(fā)突顯。Gartner在報(bào)告中指出,“數(shù)據(jù)分類為治理和安全目的的數(shù)據(jù)保護(hù)過程提供了便利。這些目的可能涵蓋從價(jià)值、訪問權(quán)限、隱私管理、存儲(chǔ)、倫理和質(zhì)量到數(shù)據(jù)保留等方面。中國的數(shù)據(jù)安全監(jiān)管要求使得數(shù)據(jù)分類成為數(shù)據(jù)安全、數(shù)據(jù)治理和數(shù)據(jù)合規(guī)項(xiàng)目的重要一步。數(shù)據(jù)分類可以幫助組織區(qū)分?jǐn)?shù)據(jù)的敏感性,提高數(shù)據(jù)安全控制的有效性?!?/span>
Data classification facilitates the process of data protection for governance and security purposes. These purposes could range from value, access rights, privacy management, storage, ethics and quality to the retention of data. China’s data security regulatory requirements make data classification a vital step for security, data governance and compliance programs. Data classification helps organizations distinguish the sensitivity of the data and improves the effectiveness of data security controls.
同時(shí),Gartner在報(bào)告中給出了用戶建議:
“通過對(duì)組織內(nèi)現(xiàn)有數(shù)據(jù)的類型、價(jià)值和敏感性進(jìn)行全面評(píng)估,確定整個(gè)組織的數(shù)據(jù)分類用例和工作。與業(yè)務(wù)部門和數(shù)據(jù)分析團(tuán)隊(duì)合作,找出數(shù)據(jù)分類至關(guān)重要的具體用例。
推行用戶培訓(xùn),并將用戶驅(qū)動(dòng)和自動(dòng)化數(shù)據(jù)分類相結(jié)合,作為數(shù)據(jù)安全治理(DSG)計(jì)劃的一部分進(jìn)行部署。
分析行業(yè)監(jiān)管機(jī)構(gòu)或國家標(biāo)準(zhǔn)委員會(huì)發(fā)布的數(shù)據(jù)分類指南和標(biāo)準(zhǔn),制定符合國家監(jiān)管要求的數(shù)據(jù)分類方案模板。
優(yōu)先選擇能與其他數(shù)據(jù)安全技術(shù)(如匿名化、加密、數(shù)據(jù)丟失預(yù)防(DLP)和數(shù)據(jù)安全平臺(tái))提供增強(qiáng)集成和互操作性的數(shù)據(jù)分類工具。此外,還應(yīng)包括增強(qiáng)的內(nèi)置分類模板和靈活的自定義標(biāo)記和標(biāo)注。
在評(píng)估基于機(jī)器學(xué)習(xí)/人工智能的數(shù)據(jù)分類工具時(shí),重點(diǎn)關(guān)注提供預(yù)訓(xùn)練和用戶訓(xùn)練解決方案的選項(xiàng),以便利用組織特定數(shù)據(jù)訓(xùn)練模型,提高準(zhǔn)確性?!?
User Recommendations
Determine organizationwide data classification use cases and efforts by conducting a thorough assessment of the types, values and sensitivity of data present within the organization. Collaborate with business departments and data analytics teams to identify specific use cases where data classification is crucial.
Implement user training and deploy a combination of user-driven and automated data classification as part of a DSG program.
Analyze data classification guidance and standards released by industry regulators or national standard committees to develop a data classification scheme that aligns with regulatory requirements in China.
Prioritize data classification tools that offer enhanced integration and interoperability with other data security technologies — such as anonymization, encryption, DLP and data security platforms. Also, other aspects include enhanced built-in categorization templates and flexible self-defined tagging and labeling.
Focus on options that provide pretrained and user-trained solutions when evaluating ML/AI-powered data classification tools to facilitate training models with organization-specific data for improved accuracy.
Gartner在報(bào)告中特別指出:
“機(jī)器學(xué)習(xí)(ML)/人工智能增強(qiáng)的數(shù)據(jù)分類工具帶來了一定程度的精度、適應(yīng)性和效率,這是以前僅通過手動(dòng)方法或基于規(guī)則的自動(dòng)化無法實(shí)現(xiàn)的。”
Machine learning (ML)/AI-enhanced data classification tools introduce a level of precision, adaptability and efficiency that were previously unattainable with manual methods or rule-based automation alone.
安恒信息作為業(yè)內(nèi)首家分類分級(jí)產(chǎn)品結(jié)合大模型進(jìn)行落地的廠商,將數(shù)據(jù)分類分級(jí)交付效率提升了30倍。利用大模型百億級(jí)別參數(shù)及海量行業(yè)數(shù)據(jù),可以覆蓋所有行業(yè)的分類分級(jí)工作,相比之前的技術(shù)手段,可以更精確識(shí)別數(shù)據(jù)含義及給出分類分級(jí)結(jié)果,并且能夠從業(yè)務(wù)視角自動(dòng)解釋分類分級(jí)結(jié)果的依據(jù)。同時(shí)能以一句話形式的自然語言指令注入專家經(jīng)驗(yàn),讓大模型按照用戶期望批量調(diào)整結(jié)果,無需再寫復(fù)雜的規(guī)則,易用性大幅提升。

與此同時(shí),AiSort分類分級(jí)工具已經(jīng)和安恒多款數(shù)據(jù)安全產(chǎn)品實(shí)現(xiàn)分類分級(jí)結(jié)果聯(lián)動(dòng),無需進(jìn)行二次開發(fā),即可實(shí)現(xiàn)精細(xì)化的數(shù)據(jù)安全監(jiān)測、預(yù)警、管控,共同構(gòu)成了一道更加嚴(yán)密和全面的數(shù)據(jù)安全防線。

參考文獻(xiàn):
Gartner, Hype Cycle for Data, Analytics and AI in China, 2024, By Ben Yan, Tong Zhang, Mike Fang, Xingyu Gu, Fay Fei, Julian Sun, Published 18 June 2024
免責(zé)聲明:
Gartner未在其報(bào)告中支持任何廠商、產(chǎn)品或服務(wù),也并不建議技術(shù)用戶只選擇有最高評(píng)分或其它特征的廠商。Gartner研究出版物代表的是Gartner研究機(jī)構(gòu)的意見,不應(yīng)解釋為對(duì)事實(shí)的陳述。Gartner對(duì)與本研究有關(guān)的所有明示或暗示的保證概不負(fù)責(zé),包括對(duì)適銷性或特定用途的適用性的任何保證。
GARTNER和HYPE CYCLE是 Gartner, Inc. 和/或其關(guān)聯(lián)公司在美國和國際上的商標(biāo)和服務(wù)標(biāo)識(shí),并在獲得許可的情況下在此使用。保留所有權(quán)利。
