WIZ State of The Cloud 2023
WIZ State of The Cloud 2023
WIZ State of The Cloud 2023
Whitepaper
Executive summary
With cloud technology constantly evolving and growing increasingly critical to business operations,
the responsibility of security professionals to stay abreast of the state of the cloud has never been
greater in order to proactively address potential threats and ensure the safe and secure deployment
of cloud solutions.
Over the past year, we have observed how cloud adoption has continued to grow with more
organizations increasing their footprint in the cloud. Many new capabilities were introduced, with the
number of possible API calls increasing by 15% in AWS, 20% in Azure, and 45% in GCP.
Although new services and their corresponding APIs expand the possibilities of how the cloud can be
utilized, they can also broaden attack surfaces and create more challenges for cloud defenders.
According to our data, 57% of companies use more than one cloud platform and therefore require
greater knowledge and expertise from their security teams who need visibility into multiple platforms
as well as the interfaces between them.
Besides novel cloud risks, well-known prevalent risks such as data exposure are also of concern. For
instance, our data shows that 47% of companies have at least one database or storage bucket
publicly exposed to the internet, and an attacker can discover and access an exposed bucket with a
guessable name (e.g. “wiz-backup”) in less than 13 hours.
In this data-driven report, based on our scanning of over 200,000 cloud accounts, including more
than 30% of the Fortune 100 environments, we analyze the latest industry trends and developments,
presenting a factual and data-based assessment of the current state and progression of cloud
technology. We examine how the cloud has evolved over the past year and attempt to shed light on
some of the complexity of cloud environments, including aspects such as organizational usage of
multi-cloud and both managed and non-managed services. We hope this can help cloud builders
and defenders ensure they have the visibility and tools that they need to continue their cloud growth
and protect their company’s assets. In addition, we will review notable cloud threats from last year
and provide insight into the speed of compromise of misconfigured environments.
The sudden upsurge in 2021 of researchers attacking cloud vendors continued into 2022. A project
sponsored by Wiz called cloudvulndb has been collecting these incidents, in part to look for patterns.
Wiz’s researchers found several critical cross-tenant vulnerabilities in multiple cloud providers (see
AttachMe, Hell's Keychain, ExtraReplica, and more). By using the insights from the past incidents and
our researchers’ expertise, we developed the PEACH cloud isolation framework to offer guidance on
better securing not only the cloud providers, but any PaaS or SaaS solution running multi-tenant
environments.
Threat actors are becoming more proficient at attacking cloud environments. LAPSUS$–one of the
most brazen–used its access at companies to move laterally through their cloud environments. Our
guidance on mitigating the risks from this group is based on existing industry best practices but
tailored to this threat actor’s particular activity against cloud environments.
AWS’s instance metadata service version 2 (IMDSv2) gained traction in a few ways this past year. A
default deployment of an EC2 uses an older version of this service (IMDSv1) which does not mitigate
SSRF and related attacks when the host contains other vulnerabilities. Even though a more secure
version was released in 2019, the default IMDSv1 is still common (most likely for legacy support).
AWS’s GuardDuty received a new detection for when IAM role credentials are suspected to be stolen
from an EC2 via this service. This risk is still relevant, as demonstrated by the threat actor UNC2903
targeting IMDSv1. A number of vendors also made changes to allow customers to enforce
IMDSv2. Finally, malware built to run specifically inside AWS Lambda functions was discovered in the
wild. Although this concept has been known for years, threat actors only just started using it.
Cloud usage continues to grow. Companies are shifting more of their workloads from on-prem to
the cloud and both adding and expanding new and existing workloads in the cloud. According to the
2022 Q3 earnings of the top three cloud providers (AWS, Azure, and GCP), revenue increased across
each cloud provider by a minimum of 20% from Q3 in 2021.
Q3 Cloud Revenue
$25B
$20B
$15B
$10B
$5B
0
AWS Azure GCP
2021 2022
Cloud providers keep increasing their offerings and their complexity. In addition to growing in size,
cloud providers are also becoming more complex. AWS has added APIs at a steady pace, with about
40 new services and 1600 new actions per year for the past 6 years1. The yearly spikes are due to the
annual AWS re:Invent conference where they release large numbers of new features.
12,000
10,000
8,000
6,000
4,000
2,000
0
2016 2017 2018 2019 2020 2021 2022
1 The API counts data was obtained by walking the commits of botocore.
top three cloud providers by 15% for AWS, 20% for Azure, and 45% for GCP2.
17,000
16,000
15,000
14,000
13,000
12,000
11,000
10,000
9,000
8,000
7,000
6,000
5,000
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
This report is based on our scanning of over 200,000 cloud accounts (AWS, OCI, Alibaba cloud
accounts, GCP projects, Azure subscriptions) including more than 30% of the Fortune 100 as
customers.
Cloud usage
22% of companies
use three or more
platforms
q
2 Privilege counts were ac uired from https ://github.com/iann0036/iam-dataset.
shows that about 78% of customers have over 80% of their workloads in a single cloud provider. In
the following diagram we have sorted our tenants by the percent of their workloads in their largest
each building represents 10% of all companies, ranked from left to right by the size of their primary platform
8 out of 10 companies have over 80% of their workloads in a single cloud provider
100%
80%
60%
40%
20%
0%
Most cloud customers concentrate their workloads in a few large accounts. Nearly all companies
running on AWS have multiple AWS accounts, but the vast majority of these companies have a few
disproportionately large accounts alongside many smaller ones. For over 97% of customers using
AWS, the largest 5% of their accounts contain over 50% of their workloads.
In other words, although most AWS customers do not maintain a single monolithic account in the
strictest sense, they do use a handful of what might be considered monolithic accounts.
AWS is the most common platform, Azure the 2nd, and GCP the 3rd. Most workloads (in this case,
virtual machines) across all companies are running on AWS (72%), and the majority of companies
(62%) choose to place more of their workloads on AWS than on other cloud providers. In other words,
no matter how we look at it, AWS is the most common primary platform among cloud customers.
3% 1%
10%
12%
13% 27%
6 2%
72%
16%
25% 30%
41%
44%
32% 10%
Platform Platform
by checking combinations or Azure
41%
usage on the right side, while the 10% 13%
middle section shows the percentage GCP
As to which database flavors are being used, PostgreSQL, Redis, and MySQL are the most prevalent,
with over 90% of companies having at least one instance of a PostgreSQL server (whether managed
or non-managed).
SQ
L
di
s
SQ
L
SQ
L ch DB DB DB nk ed ft ry DB DB DB se ra DB le DB
re Re My MS ar o go ia lu ch hi ue s ux nt nap nd ch ac el
tg cS
e
na
m
Mo
n
Ma
r Sp mc
a
ed
s
ig
Q
sm
o
nfl um
e
Sy ss
a
Co
u Or Le
v
s i Dy Me R B Co I c Ca
Po a st Do
El
Data leaks are reported in the news every week. Attackers are aware of the value of sensitive data
and the increasing difficulties in securing it. They continuously scan the internet for exposed
databases and buckets. With the average cost of a data breach now over $5 million according to
IBM’s Cost of a Data Breach 2022 Report, eliminating this risk should be a top priority.
The risk of data exposure is shockingly common. 47% of companies have at least one database or
storage bucket exposed to the internet (either managed or non-managed), and over 20% of those
cloud environments with publicly accessible buckets have buckets that contain sensitive data.
Moreover, 13% of cloud environments have at least one publicly exposed non-managed database
server, whereas for managed databases that number is 32%.
Exposed resources are compromised within hours. In experiments we ran where we created S3
buckets with names we assumed attackers might be targeting – taking well-known company names
and adding “-backup”, “_logs”, etc. – we spotted attempts to list the contents of the S3 buckets in as
little as 13 hours. In a similar test, we created an S3 bucket with an unguessable name, but referenced
it in a commit to a public GitHub repo. Attempts at listing it occurred within 7 hours. This indicates to
us the speed with which attackers could potentially find and exfil a publicly exposed S3 bucket.
Exposed S3 bucket
referenced in
GitHub repo 7 hours
Exposed S3 bucket
with common name
13 hours
Attackers are searching for exposed resources. Exposed buckets with common names are discovered within 13
hours, while exposed buckets referenced in GitHub repos are found within 7 hours.