1/57
Looks like no tags are added yet.
5 trust services criteria (AICPA) for SOC2 reports
security
ID processing methods that allow breaches / unauth’d access
availability
ID bottlenecks that prevent data from being avail when needed
processing integrity
make sure data is valid, accurate, timely, etc
confidentiality
eval employees & processes to id POTENTIAL LEAKAGE / MISHANDLING oops caps
privacy
analyze methods of collecting, storing, and using data
EUDs (computer & end user devices)
devices that DIRECTLY interact with employees and customers at the ‘edge’ of the NW, like a desktop
different from routers / stuff used by admins or devs
NW infrastructure hardware
modem
connects to internet providers’ NW
router
manages NW traffic by connecting devices to internet, its a link between modem and switches
switches
less advanced version of a router
gateway
computer device that acts as an intermediary between NWs, TRANSFORMS data between PROTOCOLS
edge-enabled devices
allows computing and all functions like that to be CLOSER to where the request initially came from
servers
physical or virtual coordination machines
firewalls (several types)
circuit level gateways (C=SOURCE)
verify source of packet
application gateway (A=ACTUAL)
inspect actual packet
NW address translation
idk
stateful multilayer inspection
combine NW address & packet filtering
next gen
diff firewall rules for diff apps and users
bus topology
tree form vs linear form (tree branches out to each part, linear is just a straight line yk)
b/c single line, its vuln if compromised
mesh topology
numerous connections between nodes, safe b/c redundancy, but costly
ring topology
nodes connected in a circular path, data must go thru each node along its path. slow, but low risk of collision
star topology
all data passes thru a central hub (switch / server), easy to ID dmg’d cables
OSI model (what are the 7 steps? whats it for?) APSTNDP
explains how protocols work and how devices comm
steps:
application
presentation
session
transport
NW
data link
physical
encap: goes from top to bottom, decap goes from bottom to top
7 layers expanded (APSTNDP)
application
doesnt represent actual app used, this is the HTTP or FTP interface between apps that a person uses
presentation
transforms #7 data into a form that other OSI devices can read, like JPEG or MPEG or ASCII, INCLUDES ENCRYPTION!!!!!
session
allows sessions between comming devices to be maintained, incl SQL, RPC, NFS
transport
validates data integrity / sees if any data was lost, incl. TCP, UOP, SSL
NW
detect errors, add source & destination IP addresses incl. IP, IPSec
data link
data packets are formatted, add MAC address, incl ISDN, PPTP
physical
converts #2 messages into bits to be transmitted, also receives bits and translates them for them to go back up the chain
firmware
locally embedded SW that tells the HW how to run, RARELY UPDATED
3 models of cloud computing
ranked from MOST control to LEAST control
IaaS (infrastructure)
entire virtual data center, incl HW NW servers storage etc, CSP is responsible for upkeep, basically CSP does everything to maintain the servers but thats it, you do everything else
PaaS (platform)
provides tools / solutions remotely for specific purposes, like building a platform or website, or advertising
SaaS (sw)
provides an app or SW thats for specific purposes, like payroll or billing (bigger picture than PaaS / more crucial to a business)
who made COSO and what is included
made by treadway commission
incl. 5 components and 20 supporting principles (lmao)
5 COSO components
gov & culture (tone, target behaviors, reinforces importance off ERM)
strategy & objective setting (strategic planning, risk appetite!!!)
performance (prioritize risks so that objectives are met and reported to SH)
review & revision (review performance over time)
info, comms, and reporting (continual process put into place to share internal and external info throughout org)
8 components of COSO - ERM
internal environment
foundation of risk profile, maybe outsource IT
objective setting
how will outsourcing help / hinder objectives
event identification
how will adopting a CSP make ID’ing events EASIER or HARDER
risk assessment
mgmt should understand the risks of its cloud strategy
risk response
reduce, avoid, share, or accept risk?
control activities
understand how traditional cntrls are changed in a cloud environment
info & comms
understand how cloud will effect timeliness & availability of comms
monitoring
mgmt should adapt monitoring mechanisms to accommodate new complexities
ERPs
facilitates large scale integration across depts. includes AIS
AIS subsystems
TPS = transaction processing
FRS = financial reporting
MRS = mgmt reporting
these record, classify, value, period, & present transactions
only stuff that accountants do that AIS doesnt do are like bank recs and AJEs
HRMS
HR mgmt sys
connected to AIS to allow real time changes to employment related data, AIS will allocate job costs and deal with direct vs indirect costs
RPAs (robotic process automations)
programs extracting info from a user interface (like a complaint form), has a lot of noise
NLP (natural language processing)
encode, decode, interpret human language, accountants use this to interpret stuff said by execs. this is like text analysis tools
neural networks
input layer output layer and hidden layer. this one has weights, will learn from its mistakes / adjust over time
2 methods of managing system changes
waterfall method (each team works on their own thing linearly) & agile method (teams work concurrently)
auditor tests
reperform, observe, inquire, document
the 2 categories in the COSO - IC integrated framework that talk about IT: (11-13-14)
cntrl activities
principle 11
cntrls over tech
info & comms
principle 13
good / high quality data
principle 14
good comms
is the entire history of blockchain visible?
YES! so its a good audit trail
what types of controls should be focused on for blockchain stuff?
PREVENTATIVE and DETECTIVE
b/c of the amount of volume / speed
3 types of system backups (availability criteria)
full backup
quick to implement, slow to take the backup
differential backup
kind of quick to implement, involves the most recent full backup & recent single differential backup, but isnt super fast at taking the backup. saves ALL data since most recent FULL backup\
incremental backup
slow to implement but quick to generate, backs up only the data since the most recent backup of any kind
types of environments
testing
devs test / debug code
development
code is written
staging
final stages of testing
production
app is made avail to end users
disaster recov
obv
baseline configs
allows IT teams to better understand root causes of issues & eval effectiveness of changes
closed loop verification
continuously monitor output from the changes & compare to the desired outcomes
acceptance criteria for change mgmt
can be qualitative or quant, should allow objective evaluation
logging
lets firms determine if policies are followed and investigate violations
waterfall method
diff teams perform separate tasks in a sequence, one after another. reqs lots of time, benefits not realized until completion. 0 customer input, idle employees
agile method
addresses issues of waterfall method, more flexible. cross functional teams working on customers needs, simultaneously. evolving final outcome
system conversion methods
direct changeover method
stop old sys and start new sys immediately. fast, but risky if sys doesnt work
parallel
old & new overlap for a bit, reqs a lot of effort
pilot
small scale pilot sys to test new sys. adj can be made prior to rollout
phased
gradual or modular, good for firms w/ multiple locations
change advisory board
helps plan for changes obviously, this is what you should do / implement if youre struggling with making changes
ways to test a sys in development to see if its working
unit testing
examine smallest increments
integration testing
done after unit testing to help ensure diff units will work tg cohesively
sys testing
verifies all modules work tg correctly / overall functionality
acceptance (beta) testing
QA phase, make sure it meets end user reqs
data life cycle 8 steps (intuitive)
definition (needs of the firm)
capture / creation (make the data)
3 challenges of external data = safety, integrity, copyrights
preparation (prep it and complete it)
make sure its up to date and hasnt been tampered with
synthesis (optional calcs can be added)
analytics & usage (INTERNAL use)
publication (external use)
archival (move it somewhere passive after you dont need it)
purging (delete)
ODS
operational data storage, repository of data from multiple sources, usually an interim area between source & warehouse
data warehouse
predefined schema
used for REPORTING & ANALYSIS (updated frequently), pulls data direct from an ODS or enterprise sys
data mart
data warehouse but more FOCUSED on a specific purpose
data lake
no predefined schema, has structured & unstructured data, i.e. data in its raw form
relational database
more complex than flat files, assures data is complete and NOT redundant
helps with business rule enforcement & comms / integration
attributes = col;umns, records = rows (alphabetical)
composite primary key
if a table doesnt have a single unique identifier, 2+ attributes can combine to uniquely ID the table
normalization
reduces database redundancy & anomalies
3 forms of normalization
1NF
each cell / field must contain only 1 piece of info
each record must be uniquely ID’d (PK)
2NF
useful for tables with composite PKs
each additional nonkey attribute must depend on ALL pks
3NF
reqs that none of the nonkey attributes depend on other nonkey attributes (transitive)
these are PROGRESSIVE, to be 3 you must be 2, and to be 2 you must be 1
model vs schema
model is high level, schema is actual implementation & execution
model (3 stages / types)
conceptual (table names & relationships), logical (adds PKs and FKs), physical (columns & data types)
ordered from LEAST detailed to MOST
fact table vs dimension table
fact = business measurements, like profit or sales
dimension = descriptive stuff like names & dates
SELECT
indicates which attributes the user wants to select from
FROM
tells DB which table contains the desired attributes
WHERE
like a filter, where (attribute) = (criteria)
more advanced is HAVING, like a WHERE function but has some sort of operation in front like HAVING COUNT(*)
INNER JOIN (no match = no data)
when you use a shared key between two tables to pull data from one table as it relates to another, like using customerID from the order table to find their phone # from the customer table
order DOES NOT matter
INNER JOIN (table 2)
ON table1.key = table2.key
LEFT / RIGHT JOIN (no match = still gives some data)
order DOES matter
basically takes the R/L table and then returns all associated values from the opposing table, incl null values
order of clauses
select, from, join on, where, group by, having, order by, limit
some felines just want giant helpings of lasagna
bus. process modeling (BPMN) pools vs swim lanes
pools
show how many orgs are involved in a process (selling org, customer, etc)
swimlanes
more granular, represents seg of duties
start events vs intermediate events vs end events
start
based on POOLS
only one per pool
thin circle
inter
two-lined circle
NOT req’d, exist when something changes the course of a process, like a time delay or error
end
based on pools
thick circle
usually 1, multiple = several ways to end / cut short
sequence flows
connect every object within a pool
smooth lined arrow
message flows: dotted arrows connecting pools
gateways
diamond shaped, represent a decision point