+ All Categories
Home > Internet > Threat Hunting

Threat Hunting

Date post: 06-Apr-2017
Category:
Upload: tripwire
View: 269 times
Download: 0 times
Share this document with a friend
110
Threat Hun+ng with Splunk Presenter: Ken Wes+n M.Sc, OSCP, ITPM Splunk, Security Market Specialist
Transcript

ThreatHun+ngwithSplunkPresenter:KenWes+nM.Sc,OSCP,ITPMSplunk,SecurityMarketSpecialist

Preworkfortoday

●  SetupSplunkEnterpriseSecuritySandbox●  InstallfreeSplunkonlaptop●  InstallMLToolkitapp

hEps://splunkbase.splunk.com/app/2890/

3

> [email protected]@kwestin

•  1.5yearatSplunk–SecurityStrategist•  BasedinPortland,Oregon•  20yearsintechnologyandsecurity•  M.Sc,OSCP,ITPM•  Trainedinoffensive&defensivesecurity

$whoami

Agenda•  ThreatHun+ngBasics

•  ThreatHun+ngDataSources

•  SysmonEndpointData

•  CyberKillChain

•  WalkthroughofACackScenarioUsingCoreSplunk(handson)

•  AdvancedThreatHun+ngTechniques&SecurityEssen+als

•  EnterpriseSecurityWalkthrough

•  ApplyingMachineLearningandDataSciencetoSecurity

LogInCreden+als

January,February&March hEps://54.144.69.125April,May&June hEps://52.55.68.96JulyandAugust hEps://54.164.82.160SeptemberandOctober hEps://52.23.227.212NovemberandDecember hEps://52.202.90.207

User:hunterPass:pr3dat0r

BirthMonth

Thesewon’twork…

AmIintherightplace?

Somefamiliaritywith…

●  CSIRT/SOCOpera+ons

●  GeneralunderstandingofThreatIntelligence

●  GeneralunderstandingofDNS,Proxy,andEndpointtypesofdata

7

Whatisthreathun+ng,whydoyouneedit?TheWhat?

• Threathun+ng-theactofaggressively

intercep+ng,trackingand

elimina+ngcyberadversariesasearlyaspossibleintheCyberKillChain2

8

TheWhy?

•  Threatsarehuman.Focusedandfundedadversarieswillnotbecounteredbysecurityboxesonthenetwork

alone.Threathuntersareac+velysearchingforthreatstopreventor

minimizedamage[beforeithappens]1

2CyberThreatHun+ng-SamuelAlonsoblog,Jan2016

1TheWho,What,Where,When,WhyandHowofEffec+veThreatHun+ng,SANSFeb2016

“ThreatHun,ngisnotnew,it’sjustevolving!”

ThreatHun+ngwithSplunk

10

Vs.

Search&Visualisa+on

Enrichment

Data

Automa+on

11

HumanThreatHunter

KeyBuildingBlockstoDriveThreatHun+ngMaturity

Ref:TheheWho,What,Where,When,WhyandHowofEffec+veThreatHun+ng,SANSFeb2016

Objec+ves>Hypotheses>Exper+se

“Agoodintelligenceofficercul,vatesanawarenessofwhatheorshedoesnotknow.Youneedadoseofmodestytoacknowledgeyourownignorance-evenmore,toseekoutyourignorance.Thentheharderpartcomes,tryingtodosomethingaboutit.ThisoFenrequiresanimmodestdetermina,on”HenryA.CrumptonTheArtofIntelligence:LessonsFromAlifeIntheCIA’sClandes+neService

12

SANSThreatHun+ngMaturity

13

AdHocSearch

Sta+s+calAnalysis

Visualiza+onTechniques

Aggrega+on MachineLearning/DataScience

85%55%50%48%32%

Source:SANSIR&ThreatHun+ngSummit2016

Search&Visualisa+on

Enrichment

Data

Automa+on

HumanThreatHunter

HowSplunkhelpsYouDriveThreatHun+ngMaturity

ThreatHun+ngAutoma+onIntegrated&outoftheboxautoma+ontoolingfromar+factquery,contextual“swim-laneanalysis”,anomaly&+meseriesanalysistoadvanceddatascienceleveragingmachinelearning

ThreatHun+ngDataEnrichment

Enrichdatawithcontextandthreat-intelacrossthestackor+metodiscerndeeperpaEernsorrela+onships

Search&VisualiseRela+onshipsforFasterHun+ng

Searchandcorrelatedatawhilevisuallyfusingresultsforfastercontext,analysisandinsight

Ingest&OnboardAnyThreatHun+ngMachineDataSourceEnablefastinges+onofanymachinedatathroughefficient

indexing,abigdatareal+mearchitectureand‘schemaontheread’technology

Hypotheses

AutomatedAnaly+cs

DataScience&MachineLearning

Data&IntelligenceEnrichment

DataSearch

Visualisa+on

Maturity

Hun+ngTools:InternalData

15

•  IPAddresses:threatintelligence,blacklist,whitelist,reputa+onmonitoringTools:Firewalls,proxies,SplunkStream,Bro,IDS

•  NetworkAr+factsandPaCerns:networkflow,packetcapture,ac+venetworkconnec+ons,historicnetworkconnec+ons,portsandservicesTools:SplunkStream,BroIDS,FPC,Neplow

•  DNS:ac+vity,queriesandresponses,zonetransferac+vityTools:SplunkStream,BroIDS,OpenDNS

•  Endpoint–HostAr+factsandPaCerns:users,processes,services,drivers,files,registry,hardware,memory,diskac+vity,filemonitoring:hashvalues,integritycheckingandalerts,crea+onordele+onTools:Windows/Linux,CarbonBlack,Tanium,Tripwire,Ac+veDirectory

•  VulnerabilityManagementDataTools:TripwireIP360,Qualys,Nessus

•  UserBehaviorAnaly+cs:TTPs,usermonitoring,+meofdayloca+on,HRwatchlistSplunkUBA,(Alloftheabove)

Persist,Repeat

ThreatIntelligence

Access/Iden+ty

Endpoint

Network

AEacker,knowrelay/C2sites,infectedsites,IOC,aEack/campaignintentandaEribu+on

Wheretheywentto,whotalkedtowhom,aEacktransmiEed,abnormaltraffic,malwaredownload

Whatprocessisrunning(malicious,abnormal,etc.)Processowner,registrymods,aEack/malwarear+facts,patchinglevel,aEacksuscep+bility

Accesslevel,privilegedusers,likelihoodofinfec+on,wheretheymightbeinkillchain

•  Third-partythreatintel•  Open-sourceblacklist•  Internalthreatintelligence

•  Firewall,IDS,IPS•  DNS•  Email

•  Endpoint(AV/IPS/FW)•  Malwaredetec+on•  PCLM

•  DHCP•  OSlogs•  Patching

•  Ac+veDirectory•  LDAP•  CMDB

•  Opera+ngsystem•  Database•  VPN,AAA,SSO

TypicalDataSources

•  Webproxy•  NetFlow•  Network

Endpoint:MicrosovSysmonPrimer

17

●  TAAvailableontheAppStore●  GreatBlogPosttogetyoustarted

●  IncreasesthefidelityofMicrosovLogging

BlogPost:hEp://blogs.splunk.com/2014/11/24/monitoring-network-traffic-with-sysmon-and-splunk/

LogInCreden+als

January,February&March hEps://54.144.69.125

April,May&June hEps://52.55.68.96

JulyandAugust hEps://54.164.82.160

SeptemberandOctober hEps://52.23.227.212

NovemberandDecember hEps://52.202.90.207

User:hunterPass:pr3dat0r

BirthMonth

SysmonEventTags

19

MapsNetworkCommtoprocess_id

Process_idcrea+onandmappingtoparentprocess_id

sourcetype=X*|searchtag=communicate

20

sourcetype=X*|deduptag|searchtag=process

21

DataSourceMapping

DemoStory-KillChainFrameworkSuccessfulbruteforce–downloadsensi+vepdfdocument

WeaponizethepdffilewithZeusMalware

Convincingemailsentwithweaponizedpdf

Vulnerablepdfreaderexploitedbymalware.Droppercreatedonmachine

Dropperretrievesandinstallsthemalware

Persistenceviaregularoutboundcomm

DataExfiltra+on

Source:LockheedMar,n

Servers

Storage

DesktopsEmail Web

Transac+onRecords

NetworkFlows

DHCP/DNS

HypervisorCustomApps

The image cannot be displayed. Your computer

PhysicalAccess

Badges

ThreatIntelligence

Mobile

CMDB

The image cannot be displ

IntrusionDetec+on

Firewall

DataLossPreven+on

An+-Malware

VulnerabilityScans

Tradi+onal

Authen+ca+on

StreamInves+ga+ons–chooseyourdatawisely

24

APTTransac+onFlowAcrossDataSources

25

hEp(proxy)sessiontocommand&controlserver

RemotecontrolStealdataPersistincompanyRentasbotnet

Proxy

ConductBusiness

Createaddi+onalenvironment

GainAccesstosystemTransac+on

ThreatIntelligence

Endpoint

NetworkEmail,Proxy,DNS,andWeb

DataSources

.pdf

.pdfexecutes&unpacksmalwareoverwri+ngandrunning“allowed”programs

Svchost.exe(malware)

Calc.exe(dropper)

AEackerhackswebsiteSteals.pdffiles

WebPortal.pdf

AEackercreatesmalware,embedin.pdf,

emailstothetarget

MAIL

Reademail,openaEachment

OurInves+ga+onbeginsbydetec+nghighriskcommunica+onsthroughtheproxy,attheendpoint,andevenaDNScall.

index=zeus_demo3

26

insearch:

Tobeginourinves+ga+on,wewillstartwithaquicksearchtofamiliarizeourselveswiththedatasources.

Inthisdemoenvironment,wehaveavarietyofsecurityrelevantdataincluding…

WebDNSProxyFirewallEndpointEmailClick

Takealookattheendpointdatasource.WeareusingtheMicrosovSysmonTA.

Wehaveendpointvisibilityintoallnetworkcommunica+onandcanmapeachconnec+onbacktoaprocess.

}Wealsohavedetailedinfooneachprocessandcanmapitbacktotheuserandparentprocess.}

Letsgetourdaystartedbylookingusingthreatinteltopriori+zeoureffortsandfocusoncommunica+onwithknownhighrisken++es.

Wehavemul+plesourceIPscommunica+ngtohighrisken++esiden+fiedbythese2threatsources.

Weareseeinghighriskcommunica+onfrommul+pledatasources.

Weseemul+plethreatintelrelatedeventsacrossmul+plesourcetypesassociatedwiththeIPAddressofChrisGilbert.Let’stakecloserlookattheIPAddress.

Wecannowseetheownerofthesystem(ChrisGilbert)andthatitisn’taPIIorPCIrelatedasset,sotherearenoimmediatebusinessimplica+onsthatwouldrequireinformingagenciesorexternalcustomerswithinacertain+meframe.

Thisdashboardisbasedoneventdatathatcontainsathreatintelbasedindicatormatch(IPAddress,domain,etc.).ThedataisfurtherenrichedwithCMDBbasedAsset/iden+tyinforma+on.

Wearenowlookingatonlythreatintelrelatedac+vityfortheIPAddressassociatedwithChrisGilbertandseeac+vityspanningendpoint,proxy,andDNSdatasources.

Thesetrendlinestellaveryinteres+ngvisualstory.ItappearsthattheassetmakesaDNSqueryinvolvingathreatintelrelateddomainorIPAddress.

ScrollDo

wn

ScrolldownthedashboardtoexaminethesethreatinteleventsassociatedwiththeIPAddress.

Wethenseethreatintelrelatedendpointandproxyeventsoccurringperiodicallyandlikelycommunica+ngwithaknownZeusbotnetbasedonthethreatintelsource(zeus_c2s).

It’sworthmen+oningthatatthispointyoucouldcreatea+ckettohavesomeonere-imagethemachinetopreventfurtherdamageaswecon+nueourinves+ga+onwithinSplunk.

Withinthesamedashboard,wehaveaccesstoveryhighfidelityendpointdatathatallowsananalysttocon+nuetheinves+ga+oninaveryefficientmanner.Itisimportanttonotethatnearreal-+meaccesstothistypeofendpointdataisnotnotcommonwithinthetradi+onalSOC.

Theini+algoaloftheinves+ga+onistodeterminewhetherthiscommunica+onismaliciousorapoten+alfalseposi+ve.Expandtheendpointeventtocon+nuetheinves+ga+on.

Proxyrelatedthreatintelmatchesareimportantforhelpingustopriori+zeoureffortstowardini+a+nganinves+ga+on.Furtherinves+ga+onintotheendpointisovenvery+meconsumingandoveninvolvesmul+pleinternalhand-offstootherteamsorneedingtoaccessaddi+onalsystems.Thisencryptedproxytrafficisconcerningbecauseofthelargeamountofdata(~1.5MB)beingtransferredwhichiscommonwhendataisbeingexfiltrated.

Exfiltra+onofdataisaseriousconcernandoutboundcommunica+ontoexternalen+tythathasaknownthreatintelindicator,especiallywhenitisencryptedasinthiscase.

Letscon+nuetheinves+ga+on.

Anotherclue.Wealsoseethatsvchost.exeshouldbelocatedinaWindowssystemdirectorybutthisisbeingrunintheuserspace.Notgood.

Weimmediatelyseetheoutboundcommunica+onwith115.29.46.99viahEpsisassociatedwiththesvchost.exeprocessonthewindowsendpoint.Theprocessidis4768.Thereisagreatdealmoreinforma+onfromtheendpointasyouscrolldownsuchastheuserIDthatstartedtheprocessandtheassociatedCMDBenrichmentinforma+on.

Wehaveaworkflowac+onthatwilllinkustoaProcessExplorerdashboardandpopulateitwiththeprocessidextractedfromtheevent(4768).

ThisisastandardWindowsapp,butnotinitsusualdirectory,tellingusthatthemalwarehasagainspoofedacommonfilename.

Wealsocanseethattheparentprocessthatcreatedthissuspicuoussvchost.exeprocessiscalledcalc.exe.

ThishasbroughtustotheProcessExplorerdashboardwhichletsusviewWindowsSysmonendpointdata.

SuspectedMalware

Letscon+nuetheinves+ga+onbyexaminingtheparentprocessasthisisalmostcertainlyagenuinethreatandwearenowworkingtowardarootcause.

ThisisveryconsistentwithZeusbehavior.Theini+alexploita+ongenerallycreatesadownloaderordropperthatwillthendownloadtheZeusmalware.Itseemslikecalc.exemaybethatdownloader/dropper.

SuspectedDownloader/Dropper

Thisprocesscallsitself“svchost.exe,”acommonWindowsprocess,butthepathisnotthenormalpathforsvchost.exe.

…whichisacommontraitofmalwareaEemp+ngtoevadedetec+on.WealsoseeitmakingaDNSquery(port53)thencommunica+ngviaport443.

TheParentProcessofoursuspecteddownloader/dropperisthelegi+matePDFReaderprogram.ThiswilllikelyturnouttobethevulnerableappthatwasexploitedinthisaEack.

SuspectedDownloader/Dropper

SuspectedVulnerableAppWehaveveryquicklymovedfromthreatintelrelatednetworkandendpointac+vitytothelikelyexploita+onofavulnerableapp.Clickontheparentprocesstokeepinves+ga+ng.

WecanseethatthePDFReaderprocesshasnoiden+fiedparentandistherootoftheinfec+on.

ScrollDo

wn

Scrolldownthedashboardtoexamineac+vityrelatedtothePDFreaderprocess.

Chrisopened2nd_qtr_2014_report.pdfwhichwasanaEachmenttoanemail!

Wehaveourrootcause!Chrisopenedaweaponized.pdffilewhichcontainedtheZeusmalware.Itappearstohavebeendeliveredviaemailandwehaveaccesstoouremaillogsasoneofourimportantdatasources.Letscopythefilename2nd_qtr_2014_report.pdfandsearchabitfurthertodeterminethescopeofthiscompromise.

LetsdigaliElefurtherinto2nd_qtr_2014_report.pdftodeterminethescopeofthiscompromise.

index=zeus_demo32nd_qtr_2014_report.pdf

39

insearch:

Letssearchthoughmul+pledatasourcestoquicklygetasenseforwhoelsemayhavehavebeenexposedtothisfile.

Wewillcomebacktothewebac+vitythatcontainsreferencetothepdffilebutletsfirstlookattheemaileventtodeterminethescopeofthisapparentphishingaEack.

WehaveaccesstotheemailbodyandcanseewhythiswassuchaconvincingaEack.Thesenderapparentlyhadaccesstosensi+veinsiderknowledgeandhintedatquarterlyresults.

ThereisouraEachment.

HoldOn!That’snotourDomainName!Thespellingisclosebutit’smissinga“t”.TheaEackerlikelyregisteredadomainnamethatisveryclosetothecompanydomainhopingChriswouldnotno+ce.

ThislookstobeaverytargetedspearphishingaEackasitwassenttoonlyoneemployee(Chris).

RootCauseRecap

42

DataSources

.pdfexecutes&unpacksmalwareoverwri+ngandrunning“allowed”programs

hEp(proxy)sessiontocommand&controlserver

RemotecontrolStealdataPersistincompanyRentasbotnet

Proxy

ConductBusiness

Createaddi+onalenvironment

GainAccesstosystemTransac+on

ThreatIntelligence

Endpoint

NetworkEmail,Proxy,DNS,andWeb

.pdfSvchost.exe(malware)

Calc.exe(dropper)

AEackerhackswebsiteSteals.pdffiles

WebPortal.pdf

AEackercreatesmalware,embedin.pdf,

emailstothetarget

MAIL

Reademail,openaEachment

Weu+lizedthreatinteltodetectcommunica+onwithknownhighriskindicatorsandkickoffourinves+ga+onthenworkedbackwardthroughthekillchaintowardarootcause.

Keytothisinves+ga+veprocessistheabilitytoassociatenetworkcommunica+onswithendpointprocessdata.

Thishighvalueandveryrelevantabilitytoworkamalwarerelatedinves+ga+onthroughtorootcausetranslatesintoaverystreamlinedinves+ga+veprocesscomparedtothelegacySIEMbasedapproach.

43

Letsrevisitthesearchforaddi+onalinforma+ononthe2nd_qtr_2014-_report.pdffile.

Weunderstandthatthefilewasdeliveredviaemailandopenedattheendpoint.Whydoweseeareferencetothefileintheaccess_combined(webserver)logs?

Click

Selecttheaccess_combinedsourcetypetoinves+gatefurther.

44

Theresultsshow54.211.114.134hasaccessedthisfilefromthewebportalofbuEergames.com.

Thereisalsoaknownthreatintelassocia+onwiththesourceIPAddressdownloading(HTTPGET)thefile.

45

ClickSelecttheIPAddress,lev-click,thenselect“Newsearch”.WewouldliketounderstandwhatelsethisIPAddresshasaccessedintheenvironment.

46

That’sanabnormallylargenumberofrequestssourcedfromasingleIPAddressina~90minutewindow.

Thislookslikeascriptedac+ongiventheconstanthighrateofrequestsoverthebelowwindow.

ScrollDo

wn

Scrolldownthedashboardtoexamineotherinteres+ngfieldstofurtherinves+gate.

No+cetheGooglebotuseragentstringwhichisanotheraEempttoavoidraisingaEen+on..

47

Therequestsfrom52.211.114.134aredominatedbyrequeststotheloginpage(wp-login.php).It’sclearlynotpossibletoaEemptaloginthismany+mesinashortperiodof+me–thisisclearlyascriptedbruteforceaEack.

Aversuccessfullygainingaccesstoourwebsite,theaEackerdownloadedthepdffile,weaponizeditwiththezeusmalware,thendeliveredittoChrisGilbertasaphishingemail.

TheaEackerisalsoaccessingadminpageswhichmaybeanaEempttoestablishpersistenceviaabackdoorintothewebsite.

KillChainAnalysisAcrossDataSources

48

hEp(proxy)sessiontocommand&controlserver

RemotecontrolStealdataPersistincompanyRentasbotnet

Proxy

ConductBusiness

Createaddi+onalenvironment

GainAccesstosystemTransac+on

ThreatIntelligence

Endpoint

NetworkEmail,Proxy,DNS,andWeb

DataSources

.pdf

.pdfexecutes&unpacksmalwareoverwri+ngandrunning“allowed”programs

Svchost.exe(malware)

Calc.exe(dropper)

AEackerhackswebsiteSteals.pdffiles

WebPortal.pdf

AEackercreatesmalware,embedin.pdf,

emailstothetarget

MAIL

Reademail,openaEachment

Wecon+nuedtheinves+ga+onbypivo+ngintotheendpointdatasourceandusedaworkflowac+ontodeterminewhichprocessontheendpointwasresponsiblefortheoutboundcommunica+on.

WeBeganbyreviewingthreatintelrelatedeventsforapar+cularIPaddressandobservedDNS,Proxy,andEndpointeventsforauserinSales.

Inves+ga+oncomplete!LetsgetthisturnedovertoIncidentReponseteam.

Wetracedthesvchost.exeZeusmalwarebacktoit’sparentprocessIDwhichwasthecalc.exedownloader/dropper.

Onceourrootcauseanalysiswascomplete,weshivedoutfocusintotheweblogstodeterminethatthesensi+vepdffilewasobtainedviaabruteforceaEackagainstthecompanywebsite.

Wewereabletoseewhichfilewasopenedbythevulnerableappanddeterminedthatthemaliciousfilewasdeliveredtotheuserviaemail.

AquicksearchintothemaillogsrevealedthedetailsbehindthephishingaEackandrevealedthatthescopeofthecompromisewaslimitedtojusttheoneuser.

Wetracedcalc.exebacktothevulnerableapplica+onPDFReader.

10minBreak!

Appendix-SQLi-DNSExfilatra+on-SplunkSecurityEssen+als

SQLi

SQLInjec+on●  SQLinjec+on●  Codeinjec+on●  OScommanding●  LDAPinjec+on●  XMLinjec+on●  XPathinjec+on●  SSIinjec+on●  IMAP/SMTPinjec+on●  Bufferoverflow

ImpervaWebACacksReport,2015

TheanatomyofaSQLinjec+onaEack

SELECT * FROM users WHERE email='[email protected]' OR 1 = 1 -- ' AND password='xxx';

[email protected]' OR 1 = 1 -- '

xxx

[email protected]

1234

AnaEackermightsupply:

…andsofarthisyear…39

index=web_vulnpasswordselect

Whathavewehere?Ourlearningenvironmentconsistsof:•  Abunchofpublically-accessiblesingle

Splunkservers•  Eachwith~5.5Mevents,fromreal

environmentsbutmassaged:

•  WindowsSecurityevents•  Apachewebaccesslogs•  BroDNS&HTTP•  PaloAltotrafficlogs•  Someothervariousbits

hEps://splunkbase.splunk.com/app/1528/

SearchforpossibleSQLinjec+oninyourevents:ü  looksforpaEernsinURIqueryfieldtoseeif

anyonehasinjectedthemwithSQLstatements

ü  usestandarddevia+onsthatare2.5+mesgreaterthantheaveragelengthofyourURIqueryfield

Macrosused•  sqlinjec+on_paEern(sourcetype,uriqueryfield)•  sqlinjec+on_stats(sourcetype,uriqueryfield)

RegularExpressionFTWsqlinjec+on_rexisasearchmacro.Itcontains:(?<injec,on>(?i)select.*?from|union.*?select|\'$|delete.*?from|update.*?set|alter.*?table|([\%27|\'](%20)*=(%20)*[\%27|\'])|\w*[%27|\']or)Whichmeans:Inthestringwearegiven,lookforANYofthefollowingmatchesandputthatintothe“injec+on”field.•  AnythingcontainingSELECTfollowedbyFROM•  AnythingcontainingUNIONfollowedbySELECT•  Anythingwitha‘attheend•  AnythingcontainingDELETEfollowedbyFROM•  AnythingcontainingUPDATEfollowedbySET•  AnythingcontainingALTERfollowedbyTABLE•  A%27ORa‘andthena%20andanyamountofcharactersthena%20andthena%27ORa‘

•  Note:%27isencoded“’”and%20isencoded<space>•  Anyamountofwordcharactersfollowedbya%27ORa‘andthen“or”

Bonus:TryouttheSQLInjec+onapp!

Summary:WebaEacks/SQLinjec+on●  SQLinjec+onprovideaEackerswitheasyaccesstodata●  Detec+ngadvancedSQLinjec+onishard–useanapp!

●  UnderstandwhereSQLiishappeningonyournetworkandputastoptoit.

●  AugmentyourWAFwithenterprise-wideSplunksearches.

DNSExfiltra+on

domain=corp;user=dave;password=12345

encrypt

DNSQuery:ZG9tYWluPWNvcnA7dXNlcj1kYXZlO3Bhc3N3b3JkPTEyMzQ1DQoNCg==.aEack.com

ZG9tYWluPWNvcnA7dXNlcj1kYXZlO3Bhc3N3b3JkPTEyMzQ1DQoNCg==

DNSexfiltendstobeoverlookedwithinanoceanofDNSdata.

Let’sfixthat!

DNSexfiltra+on

FrameworkPOS:acard-stealingprogramthatexfiltratesdatafromthetarget’snetworkbytransmi�ngitasdomainnamesystem(DNS)traffic

Butthebigdifferenceisthewayhowstolendataisexfiltrated:themalwareusedDNSrequests!hEps://blog.gdatasovware.com/2014/10/23942-new-frameworkpos-

variant-exfiltrates-data-via-dns-requests

“”

…feworganiza,onsactuallykeepdetailedlogsorrecordsoftheDNStraffictraversingtheirnetworks—makingitanidealwaytosiphondatafromahackednetwork.

hEp://krebsonsecurity.com/2015/05/deconstruc+ng-the-2014-sally-beauty-breach/#more-30872

“”

DNSexfiltra+on

hEps://splunkbase.splunk.com/app/2734/

DNSexfildetec+on–tricksofthetradeü  parseURLs&complicatedTLDs(TopLevelDomain)ü  calculateShannonEntropy

Listofprovidedlookups•  ut_parse_simple(url)•  ut_parse(url,list)orut_parse_extended(url,list)•  ut_shannon(word)•  ut_countset(word,set)•  ut_suites(word,sets)•  ut_meaning(word)•  ut_bayesian(word)•  ut_levenshtein(word1,word2)

Examples•  Thedomainaaaaa.comhasaShannonEntropyscoreof1.8(verylow)•  Thedomaingoogle.comhasaShannonEntropyscoreof2.6(ratherlow)•  A00wlkj—(-a.aslkn-C.a.2.sk.esasdfasf1111)-890209uC.4.comhasaShannon

Entropyscoreof3(ratherhigh)

Layman’sdefini+on:ascorereflec+ngtherandomnessormeasureofuncertaintyofastring

ShannonEntropy

Detec+ngDataExfiltra+on

index=brosourcetype=bro_dns|`ut_parse(query)`|`ut_shannon(ut_subdomain)`|evalsublen=length(ut_subdomain)|tableut_domainut_subdomainut_shannonsublen

TIPSq  LeverageourBroDNSdataq  CalculateShannonEntropyscoresq  Calculatesubdomainlengthq  DisplayDetails

Detec+ngDataExfiltra+on

…|statscountavg(ut_shannon)asavg_shaavg(sublen)asavg_sublenstdev(sublen)asstdev_sublenbyut_domain|searchavg_sha>3avg_sublen>20stdev_sublen<2

TIPSq  LeverageourBroDNSdataq  CalculateShannonEntropyscoresq  Calculatesubdomainlengthq  Displaycount,scores,lengths,

devia+ons

Detec+ngDataExfiltra+onRESULTS•  Exfiltra+ngdatarequiresmanyDNSrequests–lookforhighcounts•  DNSexfiltra+ontomooo.comandchickenkiller.com

Summary:DNSexfiltra+on●  Exfiltra+onbyDNSandICMPisaverycommontechnique●  Manyorganiza+onsdonotanalyzeDNSac+vity–donotbelikethem!●  NoDNSlogs?NoSplunkStream?LookatFWbytecounts

SplunkSecurityEssen+als

hEps://splunkbase.splunk.com/app/3435/

Iden+fybadguysinyourenvironment:ü  45+usecasescommoninUEBAproducts,all

freeonSplunkEnterpriseü  TargetexternalaEackersandinsiderthreatü  Scalesfromsmalltomassivecompaniesü  Savefromtheapp,sendresultstoES/UBA

ThemostwidelydeployedUEBAvendorinthemarketisSplunkEnterprise,butnooneknowsit.

Solveusecasesyoucantodayforfree,thenuseSplunkUBAforadvancedMLdetec+on.

SplunkSecurityEssen+als

TimeSeriesAnalysiswithStandardDevia+on

FirstTimeSeenpoweredbystats

GeneralSecurityAnaly+csSearches

TypesofUseCases

SplunkSecurityEssen+alsDataSources

ElectronicMedicalRecords

SourceCodeRepository

●  Howdoestheappwork?–  Leveragesprimarily| stats forUEBA–  AlsoimplementsseveraladvancedSplunksearches(URLToolbox,etc.)

●  WhycallitUEBA?–  TheseusecasesareoveninUEBAtools–  2/3ofusecasebuildonabaseline,whichisahallmarkofUEBA–  1/3areadvancedanaly+csthatothervendorsshowcaseintheirUEBA

●  Howdoesitscale?–  Appautomatestheu+liza+onofhighscaletechniques–  SummaryindexingforTimeSeries,cachinginlookupforFirstTime

SplunkEnterpriseSecurity

79

SplunkEnterprise

-BigDataAnaly+csPlaporm-

SplunkEnterpriseSecurity

-SecurityAnaly+csPlaporm-

ThreatHun+ngwithSplunk

Hypotheses

AutomatedAnaly+cs

DataScience&MachineLearning

Data&IntelligenceEnrichment

DataSearch

Visualisa+on

Maturity

ThreatHun+ngDataEnrichment

ThreatHun+ngAutoma+on

Ingest&OnboardAnyThreatHun+ng

MachineDataSource

Search&VisualiseRela+onshipsforFasterHun+ng

OtherItemsToNote

ItemstoNote

Naviga+on-HowtoGetHere

Descrip+onofwhattoclickon

Click

KeySecurityIndicators(buildyourown!)

Sparklines

Editable

Variouswaystofilterdata

Malware-SpecificKSIsandReports

SecurityDomains->Endpoint->MalwareCenter

Filterable

KSIsspecifictoRisk

Riskassignedtosystem,userorother

UnderAdvancedThreat,selectRiskAnalysis

(ScrollDown)

RecentRiskAc+vity

UnderAdvancedThreat,selectRiskAnalysis

Filterable,downtoIoC

KSIsspecifictoThreat

Mostac+vethreatsource

Scrolldown… Scroll

UnderAdvancedThreat,selectThreatAc+vity

Specificsaboutrecentthreatmatches

UnderAdvancedThreat,selectThreatAc+vity

Toaddthreatintelgoto:Configure->DataEnrichment->ThreatIntelligenceDownloads

Click

Click“ThreatAr+facts”Under“AdvancedThreat”

Click

Ar+factCategories–clickdifferenttabs…

STIXfeed

Customfeed

UnderAdvancedThreat,selectThreatAr+facts

ReviewtheAdvancedThreatcontent

Click

DatafromassetframeworkConfigurableSwimlanes

Darker=moreevents

Allhappenedaroundsame+meChangeto“Today”ifneeded

AssetInves+gator,enter“192.168.56.102”

DataScience&MachineLearningInSecurity

92

Disclaimer:Iamnotadatascien+st

TypesofMachineLearningSupervisedLearning:generalizingfromlabeleddata

SupervisedMachineLearning

95

DomainName TotalCnt RiskFactor AGD SessionTime RefEntropy NullUa Outcome

yyfaimjmocdu.com 144 6.05 1 1 0 0 Maliciousjjeyd2u37an30.com 6192 5.05 0 1 0 0 Maliciouscdn4s.steelhousemedia.com 107 3 0 0 0 0 Benignlog.tagcade.com 111 2 0 1 0 0 Benigngo.vidprocess.com 170 2 0 0 0 0 Benignstatse.webtrendslive.com 310 2 0 1 0 0 Benigncdn4s.steelhousemedia.com 107 1 0 0 0 0 Benignlog.tagcade.com 111 1 0 1 0 0 Benign

UnsupervisedLearning:generalizingfromunlabeleddata

UnsupervisedMachineLearning

•  Notuning

•  Programma+callyfindstrends

•  UBAisprimarilyunsupervised

•  Rigorouslytestedforfit

97

AlgorithmRawSecurityData AutomatedClustering

98

MLToolkit&Showcase•  SplunkSupportedframeworkforbuildingMLApps

–  Getitforfree:hEp://+ny.cc/splunkmlapp

•  LeveragesPythonforScien+ficCompu+ng(PSC)add-on:–  Open-sourcePythondatascienceecosystem–  NumPy,SciPy,scitkit-learn,pandas,statsmodels

•  Showcaseusecases:PredictHardDriveFailure,ServerPowerConsump+on,Applica+onUsage,CustomerChurn&more

•  Standardalgorithmsoutofthebox:–  Supervised:Logis+cRegression,SVM,LinearRegression,RandomForest,etc.–  Unsupervised:KMeans,DBSCAN,SpectralClustering,PCA,KernelPCA,etc.

•  Implementoneof300+algorithmsbyedi+ngPythonscripts

MachineLearningToolkitDemo

100

SplunkUBA

103

SplunkEnterprise

-BigDataAnaly+csPlaporm-

SplunkEnterpriseSecurity

-SecurityAnaly+csPlaporm-

ThreatHun+ngwithSplunk

ThreatHun+ngDataEnrichment

ThreatHun+ngAutoma+on

Ingest&OnboardAnyThreatHun+ng

MachineDataSource

Search&VisualiseRela+onshipsforFasterHun+ng

Hypotheses

AutomatedAnaly+cs

DataScience&MachineLearning

Data&IntelligenceEnrichment

DataSearch

Visualisa+on

Maturity

UserBehaviorAnaly+cs

-SecurityDataSciencePlaporm-

104

MachineLearningSecurityUseCasesMachine

LearningUseCases

PolymorphicAEackAnalysis

BehavioralPeerGroupAnalysis

User&En+tyBehaviorBaseline

Entropy/RareEventDetec+on

CyberAEack/ExternalThreatDetec+on

Reconnaissance,BotnetandC&CAnalysis

LateralMovementAnalysis

Sta+s+calAnalysis

DataExfiltra+onModels

IPReputa+onAnalysis

InsiderThreatDetec+on

User/DeviceDynamicFingerprin+ng

SplunkUBAUseCases

ACCOUNTTAKEOVER•  Privilegedaccountcompromise•  Dataexfiltra+on

LATERALMOVEMENT

•  Pass-the-hashkillchain•  Privilegeescala+onSUSPICIOUSACTIVITY•  Misuseofcreden+als•  Geo-loca+onanomalies

MALWAREATTACKS•  Hiddenmalwareac+vityBOTNET,COMMAND&CONTROL

•  Malwarebeaconing•  Dataleakage

USER&ENTITYBEHAVIORANALYTICS•  Suspiciousbehaviorbyaccountsor

devices

EXTERNALTHREATSINSIDERTHREATS

SplunkUserBehaviorAnaly+cs(UBA)•  ~100%ofbreachesinvolvevalidcreden+als(MandiantReport)•  Needtounderstandnormal&anomalousbehaviorsforALLusers•  UBAdetectsAdvancedCyberaEacksandMaliciousInsiderThreats•  LotsofMLunderthehood:

–  BehaviorBaselining&Modeling–  AnomalyDetec+on(30+models)–  AdvancedThreatDetec+on

•  E.g.,DataExfilThreat:–  “Sawthisstrangelogin&datatransferforuserkwes+n

at3aminChina…”–  SurfacethreattoSOCAnalysts

Raw Events

1

Statistical methods

Security semantics

2 Threat Models

Lateralmovement

ML

Patterns

Sequences

Beaconing

Land-speedviola+on

Threats

Kill chain sequence

5

Supporting evidence

Threat scoring

Graph Mining

4

Con

tinuo

us s

elf-l

earn

ing!

Anomalies graph

Entity relationship graph

3

Anomalies

RAW SECURITY EVENTS

ANOMALIES ANOMALY CHAINS (THREATS)

MACHINE LEARNING

GRAPH MINING

THREAT MODELS

Lateral Movement Beaconing Land-Speed Violation

HCI

Anomalies graph Entity relationship graph

Kill chain sequence Forensic artifacts Threat/Risk scoring

FEEDBACK

SplunkUBADemo

108

SecurityWorkshops

●  SecurityReadinessWorkshop●  DataScienceWorkshop●  EnterpriseSecurityBenchmarkAssessment

SecurityWorkshopSurvey

hCps://www.surveymonkey.com/r/3T6T9TH

[email protected]:@kwes+nlinkedin.com/in/kwes+n


Recommended