《Splunk：2025数据管理新规则：AI时代下的价值创造研究报告（英文版）（20页）.pdf》由会员分享，可在线阅读，更多相关《Splunk：2025数据管理新规则：AI时代下的价值创造研究报告（英文版）（20页）.pdf（20页珍藏版）》请在三个皮匠报告上搜索。

1、The New Rules of Data ManagementCreating value in the AI eraContents3Introduction:Finding value in the data clutter4Chapter1:Data at the crossroads7Chapter2:Data management practices have fallen behind9Chapter3:The new rules of data management 12Chapter4:Leaders play by the new rules15Chapter5:The s

2、ymbiotic relationship of data management and AI17Conclusion:Getting your data house in order19 Methodology20 About SplunkThe New Rules of Data Management|Splunk3Finding value in the data clutterYou know the paradox:drowning in data but starving for insights.The adage was never more true than it is t

3、oday.The right data fuels insights that help organizations invent better customer experiences,identify malicious threats,and improve countless other processes to strengthen digital resilience.The plain fact,though,is that cloud services,connected devices,and AI are overwhelming organizations.And ins

4、tead of thoughtfully arranging their data,they are stockpiling it like a garage cluttered with gardening tools,camping gear,and childhood memorabilia.We wanted to know how organizations are cleaning out their data garages(so to speak),so we surveyed 1,475 IT,engineering,and cybersecurity professiona

5、ls across the globe about their data management practices.Weve based this report on our findings,revealing the best practices to ensure data is on hand when you need it,while creating more value.Organizations have long followed the conventional wisdom of centralizing data into one place to unify vis

6、ibility and better make sense of it.Although this practice offered organizations some control and visibility,data structures became more complex.Consequently,data management became more difficult,requiring strategies that went beyond simply centralizing data into one location.In an attempt to contro

7、l costs and manage the explosion of data,organizations started expanding their storage locations with a medley of hybrid environments,opening the door for new sets of challenges.We think theres a better way.The new rules of data management can help you realize your security and observability objecti

8、ves and advance your mission,while you also optimize costs and compliance.Keep reading to see what data management leaders do differently.Discover how to tamp down data complexity and maximize its value in the AI era.INTRODUCTIONThe New Rules of Data Management|Splunk4Data at the crossroadsCHAPTER 1

9、Whatsstandinginthewayofyourdatamanagementstrategy?Data security and compliance69%67%41%35%30%28%26%Data volume and growthDefining data tiersCost managementData collectionData migrationAccess and retrieval speedThe survey confirms what many organizations may have suspected for years the exponential r

10、ise of data is giving way to increased complexity that makes it more difficult to access,analyze,and secure data,as well as comply with regulatory mandates.This is why having a sound and comprehensive data management strategy is crucial for digital resilience.But here again,volume and too many siloe

11、d data stores get in the way.In fact,67%of survey respondents cite data volumes and growth as a challenge when implementing their data strategy,surpassed only by 69%who call maintaining data security and compliance a top data management obstacle.They agree that defining data tiers,cost management,an

12、d other activities were also obstacles.The New Rules of Data Management|Splunk5TherealworldconsequencesofdatamanagementchallengesPoor decision-making9%20%31%40%Failure to meet compliance mandates11%27%29%33%Competitive disadvantage9%45%38%8%Unplanned downtime12%45%39%4%Poor customer service/experien

13、ce19%39%37%5%Organizations wrestling with these data management issues are also feeling far-reaching business impacts.Sixty-two percent of respondents claim that difficulties with data management resulted in compliance failures(33%significant impact,29%moderate impact),71%say they led to poor decisi

14、on-making(40%significant impact,31%moderate impact),and 46%confirm they led to competitive disadvantages(8%significant impact,38%moderate impact).Data redundancy is also a serious dilemma for organizations trying to stay afloat in a tsunami of data.Fifty-nine percent of respondents reveal their curr

15、ent data management strategy has somewhat worsened the rate of data duplication,and 20%say the problem is significantly worse.Did not experienceModerate negative impactSignificant negative impactMinimal negative impactThe New Rules of Data Management|Splunk6Breaking down the cost of data managementD

16、ata management costs are on the rise for almost everyone.Ninety-one percent of respondents reveal they spent more on data management this year than in the previous year.Respondents call out volume and compliance again,this time for driving increased costs nearly three-quarters(73%)label data volume

17、as a primary cause,and shifting compliance regulations came in second at 71%.The latter reflects a groundswell of more expansive and rigorous compliance mandates,requiring organizations to understand exactly where and how data is stored and protected across their ecosystem,and more importantly,who h

18、as access to data.Complying with current regulations such as FedRAMP,ISO27001,PCI,and HIPAA(just to name a few)require more financial investment now because more is at stake.Organizations risk potential fines and importantly,collateral damage to their reputations and customers trust if they fail to

19、comply.And even long-standing regulations have become more expansive and demanding.The European Unions General Data Protection Regulation(GDPR),for example,requires comprehensive visibility across an organizations entire data and customer environment,and is likely one of the costliest regulations to

20、 support from a data management perspective.When evaluating budget allocation relative to the data lifecycle,respondents report spending 6%less on storage and 7%less on indexing on average.In light of steadily rising data costs,organizations have sought out less expensive options for data storage.Wh

21、ile convenient,without the right overarching strategy and controls,these distributed storage methods spread across multiple clouds,data lakes,and other storage locations risk duplication,redundancy,and governance issues.Whats more,adopting multiple storage options can create unintended complexity th

22、at may thwart efforts to understand and streamline costs more than a quarter(26%)of respondents maintain they arent able to accurately calculate the ROI of their data management investment.On average,respondents admit they spent more than a quarter(28%)of their data management budget this year on se

23、arch and analysis,up slightly from 24%spent the previous year.This suggests a keen interest in not only reducing the noise in data,but also mining its value.reporttheiroverallspendondatamanagementhasincreasedcomparedtothepreviousyear91%Thetopdriversofincreaseddatamanagementcosts73%Increasing data vo

24、lumes71%Shifting regulations62%Data management technology 53%Data security needs 14%Sustainability effortsThe New Rules of Data Management|Splunk7Data management practices have fallen behindA data management strategy is a set of practices to help organizations tame data complexity and manage its lif

25、ecycle.However,many organizations havent evolved their data management practices in line with data growth and complexity.The old way of data management either requires your data to be consolidated into one location at significant cost,or that you live with data silos and sacrifice visibility.As a re

26、sult,organizations are compelled to migrate data frequently and struggle with privacy across their environments.Data access is a prevailing issue.Fifty-three percent admit they have to log into different platforms to access different data sources.And few respondents say their data management strateg

27、y includes components such as unified visibility(13%)and unified accessibility(11%).One reason for this disparity could be the need to break down organizational silos across multiple systems and teams that makes achieving unified visibility and unified accessibility difficult to realize.Many organiz

28、ations are still moving data from disparate sources to consolidate their environments and gain visibility.Forty-seven percent move data monthly,and most say they have been migrating more data to cloud infrastructure(76%)over the last two years.But moving data monthly carries risk,opening the door to

29、 potential security breaches,data leaks,and compliance violations if not properly handled.Depending on the size and complexity of the data being moved,costs can quickly add up and take a toll on an organizations bottom line.CHAPTER 2ThemanypracticesthatmakeupadatamanagementstrategyData lifecycle man

30、agementData security and complianceData tieringData pipeline managementData quality(e.g.accuracy,timelines,validity)Real-time data processing and streamingData reuse for security and observabilityData virtualizationUnified accessibilityData documentationUnified visibilityAutomated data integration75

31、%49%36%73%48%24%16%13%11%14%13%8%The New Rules of Data Management|Splunk8And while organizations may have a data management strategy in place,many struggle with fundamental governance or enforcement.A considerable portion of respondents reveal the following data management policies are not well-enfo

32、rced:role-based data access(57%),instructions about where data types should be stored(57%),and defined data retention periods(44%).These loopholes can jeopardize compliance standing,potentially resulting in hefty fines,legal repercussions,loss of brand and reputation,and diminished customer trust,am

33、ong other ramifications.Additionally,79%of respondents dont have a policy governing data destruction(33%have no plans to create one)further muddying the waters.Organizations are already overwhelmed by a complex and noisy data landscape.Inconsistent and outmoded data management practices only add to

34、their struggles.OrganizationswaffleonenforcingpoliciesA defined retention period for how long data should be stored in each locationA clear policy that instructs where specific data types should be storedA clear data access policy based on employee roleA clear policy that governs the destruction of

35、data1%33%8%46%44%17%47%1%9%57%33%2%8%57%33%4%No,and dont have a plan to create oneYes,but this policy is not well-enforcedYes,and this policy is well-enforcedNo,but are working towards building oneThe New Rules of Data Management|Splunk9The new rules of data management Organizations with a forward-t

36、hinking data management strategy are primed for digital resilience,as they can access and process data faster,and have higher-quality data to surface insights and produce more reliable outcomes.They have adopted practices that provide a clear upside for data management.According to the survey,the tw

37、o practices accounting for the lions share of organizations data management strategies are data lifecycle management(75%)and data pipeline management(73%).(Well come back to these foundational practices in the next chapter.)Looking more closely at the data reveals other practices that,while less uti

38、lized today,move respondents a step closer to value creation.Data quality,data reuse,data tiering,and data federation all help organizations access,see,and understand their most critical information.In short,these practices help organizations know what data is being generated in their enterprise and

39、 allow them to access it cost effectively,regardless of where it resides.Data qualityData quality is a significant part of a data management strategy for 48%of respondents,who claim theyve experienced a myriad of improvements compared to those who dont emphasize it.For example,73%of organizations th

40、at make data quality a priority(vs.51%of all other respondents)say mean time to respond(MTTR)has improved.Theyre also more likely to successfully neutralize threats(54%vs.41%all other respondents),identify root causes(45%vs.34%all other respondents),and improve threat detection capabilities(61%vs 37

41、%).Data reuseEach data source can serve multiple purposes,but factors like data accessibility and proprietary formatting often drive data duplication and blind spots.Data reuse might not be as pervasive as other practices only about 16%of respondents say data reuse for security and observability com

42、prises their data management strategy.But if anything,this indicates more opportunities to save costs by avoiding redundant data collection,enhancing collaboration,engaging in data stewardship across teams,and generating new insights by combining datasets from different sources.Organizations that in

43、clude data reuse in their data management strategy say it has generated notable value.Among other benefits,they are less likely to face hurdles when handling high volumes of data(46%vs.71%all other respondents).Theyre also more likely to reduce the impact of incidents(52%vs.35%all other respondents)

44、,and experience fewer data breaches(44%vs.33%all other respondents).Organizations that reuse their data also see better threat detection performance(62%vs.47%all other respondents).Data tieringData tiering prioritizes data based on factors such as access frequency,age of the data,and usage patterns.

45、According to the survey,36%of organizations employ this practice as a part of their data management strategy,saying they aim to reduce data storage costs and accelerate access times for commonly used data types.CHAPTER 3The New Rules of Data Management|Splunk10Data tiering enhances access times and

46、security while reducing costsBenefits ranked number one by respondentsReduced storage costsIncreased security for older data typesIncreased data analysis productivityAccelerated access times for commonly used data types50%32%10%8%The New Rules of Data Management|Splunk11Respondents who have implemen

47、ted data tiering experience many benefits,with 50%ranking reduced storage costs as the number one positive impact,followed by accelerated access for commonly used data types(32%),increased security for older data types(10%),and increased data analysis productivity(8%).Organizations that tier data ar

48、e also less likely to encounter challenges with access and retrieval speed(18%vs.31%all other respondents),cost management(18%vs.44%all other respondents),and data migration(18%vs.34%all other respondents).Data federationOrganizations that employ federation can access and analyze data from multiple,

49、disparate data sources and locations as though it were a single dataset and without moving the data.However,few have mastered the art.While 92%confirm having some form of a federated practice,only 20%claim its fully implemented.With so many storage locations,access methods,analytics platforms,and da

50、ta workflows to navigate,a federated data management strategy makes a lot of sense.Organizations that have adopted data federation,whether fully or partially,reveal a slew of benefits,including faster data access(67%),improved data governance(54%),and improved compliance posture(47%).The survey show

51、s that a federated data management strategy provides organizations enormous advantages across security,observability,AI,and other critical areas.Ultimately,if organizations aim to maximize the value of their data,adding federation to complement their current data management strategy will be key.What

52、datafederationcandoforyou67%Faster data access54%Improved data governance47%Improved compliance posture37%Reduced data redundancy36%Cost savings21%Minimized data movement The New Rules of Data Management|Splunk12Leaders play by the new rulesCHAPTER 4DatamanagementleadersslashcostsPercentage of respo

53、ndents who reported cost savings62%34%LeadersOthersWhen developing a winning data management strategy,the survey indicates organizations that have adopted a trifecta of practices fully implemented data federation,data pipeline management,and data lifecycle management are often ahead of their peers.T

54、hese data management leaders not only make strategic data management investments,they also realize a host of business benefits.The leader cohort reports greater business performance improvement over the last two years in several key areas compared to all other respondents,including net operating pro

55、fit margin(69%vs.56%),sustainability(58%vs.38%),and speed of innovation(55%vs.44%).Data management leaders are also more likely to state their data management strategy has enhanced other key data-related metrics,such as speed to access(79%vs.73%of all other respondents),speed of overall data process

56、ing(76%vs.69%of all other respondents)and their amount of computational overhead(62%vs.45%all other respondents).The leader cohort also sees other valuable benefits from their data management strategy,most significantly cost savings(62%vs 34%all other respondents).The New Rules of Data Management|Sp

57、lunk13Modern data management strengthens cybersecurityA modern data management strategy does more than wrangle and organize data;it also has a measurable impact that boosts other security outcomes.Leaders report significant improvement in all aspects of TDIR threat detection(26%vs.12%all other respo

58、ndents),investigations(22%vs.9%all other respondents),and response(33%vs.20%all other respondents).Data complexity can expand an organizations attack surface,providing more opportunities for threat actors to engage in nefarious activities.Left unchecked,this complexity can impede business success.Ho

59、wever,a winning data management strategy helps organizations align their data with security goals.Data management leaders report faster mean time to respond(MTTR)(79%vs.61%all other respondents),more successful threat neutralizations(65%vs.45%all other respondents),quicker root cause identification(

60、47%vs.38%all other respondents),and fewer breaches(43%vs.34%all other respondents).DatamanagementleaderselevatesecuritypostureAreas that have slightly or significantly improvedLeadersOthersMTTR79%61%Frequency of successful threat neutralizations65%45%Root cause identification47%38%Total number of da

61、ta breach incidents43%34%The New Rules of Data Management|Splunk14Robust data management boosts observability,ITOps practicesA data management strategy composed of fully-implemented data federation,data pipeline management,and data lifecycle management has similarly rewarding outcomes in ITOps and o

62、bservability practices.In observability,data management leaders experience substantial gains in scalable observability model building(79%vs.60%all other respondents).Leaders also confirm their data strategy improved performance optimization for app infrastructure(79%vs.60%all other respondents),as w

63、ell as critical business process monitoring(76%vs.58%all other respondents).As in security,the winning trifecta of data management practices improves IT metrics for leaders.Leaders see significant gains in KPIs such as mean time to resolve(MTTR)incidents(78%vs 58%all other respondents)and log volume

64、 and pattern optimization(56%vs 38%all other respondents).DataleadersboostobservabilityoutcomesAreas that have slightly or significantly improvedLeadersOthersScalable observability model buildingPerformance optimization for app and infrastructureCritical business processes monitoringIncident respons

65、e and root cause analysis79%79%76%54%60%60%58%51%The New Rules of Data Management|Splunk15The symbiotic relationship of data management and AIThe survey suggests the relationship between data management and AI is mutually beneficial.AI depends on quality data,so a strong data management strategy pla

66、ys a vital role in how AI models perform.The inverse is also true AI helps fill in the gaps of organizations data management practices by boosting productivity and automation when woven into workflows.A strong data management strategy will be a force multiplier for AI implementation.Across the board

67、,survey respondents hail the benefits of their data management strategy on AI,with 85%saying it provides AI with enough data volume and variety to generate valuable insights(41%strongly agree,44%somewhat agree).Additionally,74%report their data management strategy removes bias from the datasets from

68、 which AI models learn(37%strongly agree,37%somewhat agree).And 82%say their organizations data strategy has improved the accuracy of their machine learning models(38%strongly agree,44%somewhat agree)all of which lay the groundwork for competitive advantage as they build out their AI implementations

69、 in a crowded market.Whats more,81%of organizations also say they leverage insights from security and observability tools to enhance AI model training and performance(39%strongly agree,42%somewhat agree).While AI offers many advantages for data management and vice versa AI also introduces new obstac

70、les that continue to be a source of frustration.Survey respondents cite that AI has made data integration harder and contributed to the existing challenge of high data volumes.CHAPTER 5N/ASomewhat disagreeSomewhat agreeStrongly agreeStrongly disagreeAIsuccessstartswithdatamanagementOur data strategy

71、 provides AI with the volume and variety of data needed to drive insights.12%44%41%1%Our organization uses insights from security and/or observability tools to enhance AI performance.14%42%39%3%2%Our data strategy removes bias from the datasets our AI learns from.Our data strategy has improved the a

72、ccuracy of our machine learning models.14%44%38%2%2%4%8%14%37%37%2%The New Rules of Data Management|Splunk16Yet,despite these obstacles,AI has a largely positive effect on data management.Virtually all respondents(98%)agree that AI made their data management strategy easier(33%say significantly easi

73、er).AI delivered the most value by performing routine,administrative functions 73%of respondents believe AI enhanced data quality by automating repetitive tasks.AI also opened up new opportunities.Fifty-nine percent state it helped with data discovery,including scanning large datasets to identify pa

74、tterns,trends,and anomalies.The value of data management and AI are interwoven,with each enhancing the effectiveness of the other.The quality and accuracy of your AI is directly related to the data it has access to and the quality of that data.Conversely,AI enhances data management by automating pro

75、cesses,improving security,and optimizing storage,creating a cycle of continuous enhancement.Organizations that leverage both effectively realize a significant competitive advantage in data-driven decision-making.Enhancedataqualityviaautomatingrepetitivetasks73%Easierdatadiscoverytoidentifypatterns,t

76、rends,andanomalies59%How AI is transforming data management for the betterThe New Rules of Data Management|Splunk17Getting your data house in orderLike any spring cleaning project to reorganize your drawers,closets,and garage,restructuring your approach to managing data is an opportunity to reset.It

77、 helps you not only declutter,but also make room for new possibilities.But first,youll need to start with the basics:Know what data your organization generates and prioritize business goals and use cases.Here are a few recommendations to help you maximize your datas value from the ground up.1.Know y

78、our data and classify it To lay the foundation of data management,you must first understand the data being generated in your organization.Then,define your target use cases according to how data will be used(real-time detection vs.historical investigation),relative to the business constraints(retenti

79、on requirements,for example).From there,you can then identify which data management practices can help you meet those needs.Classifying your data will also require a strong data governance policy,along with data retention and role-based access.So make sure youre providing regular policy training to

80、your teams so they understand where the data lives and how it can and should be used.2.Keep your data cleanQuality matters.That holds especially true for your data.However,only half of survey respondents prioritize data quality as a core component of their data management strategy.Even if your data

81、is federated,accessible,and indexed,your data management strategy can still fail.Why?Because you dont have the right data powering your systems,processes,platforms,and applications.Having the right data is an iterative process that starts from the moment it is generated.It should be fundamentally ac

82、curate,complete,and formatted to meet needs as they arise,ensuring its optimized to create value.Prioritizing quality data will be especially important when you start implementing AI.(Remember the old adage,“Garbage in,garbage out?”)Good,clean data will help your AI models perform more accurately an

83、d give you better outcomes.3.Access your data without moving itWe get it,you need to have a single source of truth,and that means ensuring all your data is accessible.Thats where a data federation practice provides the most value,offering unimpeded access to all of your data,regardless of where it l

84、ives and without costly migrations.The ability to access your data at rest is critical.Its especially important when accessing data stored in diverse locations,necessary for making informed business decisions.For example,when a user requires additional information during the threat hunting process,t

85、hey need the ability to run ad-hoc searches against the external data store where that data resides to gather insights and make the right decisions.Data federation enables you to easily reach for specific data related to an incident,allowing you to make accurate,and better informed decisions about y

86、our current environment,and how to keep your systems protected in the future.4.Take a platform approach to your dataWhile you might be able to query your data from a number of separate tools,you will still need to unify your data so you can clearly see the entire picture.Implementing a unified data

87、platform one fully equipped with federation capabilities that deliver unified accessibility without having to move data at all or log into different platforms will not only bring your data into full view,but also make it easy to locate and use for any use case,without breaking the bank.In addition t

88、o a holistic view of your data,a unified data platform will also help pare down multiple or redundant tools,streamline workflows,reduce integration headaches from multiple vendors,and ease“swivel chair syndrome”and other issues.Whether youre leveraging data for security or observability,or both(thin

89、k data reuse),or using it to drive AI,a data platform that enables pipeline management,analytics,and federated access helps you serve the right data to the right teams.CONCLUSIONThe New Rules of Data Management|Splunk1718The New Rules of Data Management|SplunkRedefine your data management strategy w

90、ith SplunkPerspectivesbySplunkbyleaders,forleadersFind out how executives and business leaders are rethinking their data management strategy to address industry challenges and realize new opportunities across security,observability,and AI.TheCISOReportDiscover how CISOs and their boards are bridging

91、 critical gaps on top priorities,including collaboration,data reuse,compliance approaches,and success metrics.Get executive insightsDownload the reportThe New Rules of Data Management|Splunk19Methodology Oxford Economics researchers surveyed 1,475 IT,engineering,and cybersecurity professionals from

92、November 2024 through January 2025.Respondents were in Australia,France,Germany,India,Japan,New Zealand,Singapore,United Kingdom,and United States.They also represented 16 industries:business services,construction and engineering,consumer packaged goods,education,financial services,government(federa

93、l/national,state,and local),healthcare,life sciences,manufacturing,technology,media,oil/gas,retail/wholesale,telecom,transportation/logistics,and utilities.Respondents defined as“data management leaders”consist of organizations that have applied fully implemented data federation,data pipeline manage

94、ment,and data lifecycle management.The New Rules of Data Management|Splunk19Splunk,Splunk and Turn Data Into Doing are trademarks and registered trademarks of Splunk LLC.in the United States and other countries.All other brand names,product names or trademarks belong to their respective owners.2025

95、Splunk LLC.All rights reserved.25_CMP_report_the-new-rules-of-data-management_v11About SplunkSplunk,a Cisco company,helps make organizations more digitally resilient.Leading organizations use our unified security and observability platform to keep their digital systems secure and reliable.Organizations trust Splunk to prevent infrastructure,application,and security incidents from becoming major issues,recover faster from shocks to digital systems,and adapt quickly to new opportunities.Keep the conversation going with Splunk.