Nov 21, 2019 - Vanessa Wegner    

A brief guide to multicloud security

Five challenges and seven best practices to consider for your multicloud strategy.

Many agree that multicloud is worth the risk.

The multicloud trend has taken hold in recent years, with RightScale finding that 84% of enterprises run a multicloud strategy. With multicloud, organizations deploy applications across two or more cloud platforms, like AWS, Azure, or Google Cloud.

Increased flexibility is one of the biggest appeals of a multicloud strategy. Companies avoid vendor lock-in by deploying workloads to different cloud platforms based on cost and application needs. Hyperscale cloud vendors have data centers across the globe, so organizations are able to control their cloud expenditures by scheduling workloads based on location and local time. Multicloud also protects business operations by reducing down time, and improving resilience in the event of an outage or workload-disruptive breach (like a DDoS attack).

However, multicloud still has drawbacks that require careful consideration. The increased complexity of a multicloud environment exponentially increases an organization’s attack surface and level of risk. Most of these risks can be mitigated with a thorough assessment and strategy addressing security needs – and as a study from IDG and IBM has found, 70% of survey respondents agreed that the benefits of multicloud outweigh the risks.

That being said, there’s a lot to consider. In this blog, we’ll run through some of the top security challenges of multicloud, and dig into the strategies to conquer them. If you're short on time, feel free to skip down to the best practices section.

Guide to the cloud Harness the power of the cloud with microservices, cloud-agnostic DevOps, and workflow portability. Learn more

Key security challenges and how to manage them

Access and permissioning

multicloud adds complexity to your identity and access management efforts. Employees need access to multiple cloud services as part of their daily work, and will access your data from a multitude of locations and devices. We recommend you take a Zero Trust approach here: Allow access on an as-needed basis, and no more. Data classification levels can help you streamline access determinations across different clouds, but the key idea is that limited access will both protect your most mission critical and sensitive information, and allow you a clear view of when (and by whom) that information is accessed.

Staying up to date

While this is a security concern for any cloud use, upgrades and patching in multicloud are more challenging because the vulnerabilities and mitigations from each cloud service provider are different. multicloud complexity also makes it difficult to keep track of vulnerabilities as applications communicate across multiple clouds. Mike Bursell from RedHat calls this need “workload freshness” – and suggests that this might require you to upgrade or patch in place, restart the workload with the latest image, or check and reload recent dependencies, in order to maintain the most recent versions of any dependent libraries, middleware, or executables.

A disjointed view of security

Most cloud vendors offer native tools to help you manage security within their cloud platform, and most of those tools can’t be applied to other vendors. This disjointed approach to monitoring makes it difficult to gain a thorough understanding of all the vulnerabilities present in your infrastructure.

Instead of making piecemeal security sense, adopt a multicloud management tool that serves as a single pane of glass into all the happenings across all of your cloud platforms. Bursell notes that any monitoring tool needs to be fully aware of the scope of your deployment. It’s also important to have regular, if not real-time, updates to your data view so that you’re aware of unusual changes or activities and can address attacks as they come in. A centralized tool is also valuable for conducting forensic analysis of your systems in the event of a late-discovered breach.

Control plane complexity

RedHat’s Bursell defines the control plane as any communication which controls your applications or how they are run. In addition to securing communications between and within applications, all scheduling, monitoring, and routing communications should also be encrypted. It’s critical to secure the administration, logging, and audit functionality of your applications (lest you want to give hackers the opportunity to take down your entire infrastructure). David Locke of World Wide Technology writes that security functionality and enforcement needs to be uniform within all of your cloud environments, allowing those functions to communicate and coordinate between themselves and support security automation.

Application hardening

When hardening your infrastructure, Bursell recommends knowing what APIs are exposed, understanding what controls you have on them, and planning what mitigations you can apply if they come under attack. Tripwire notes that any software that your organization develops or acquires from a third party must be patched and security hardened by your organization.

Best practices

Need a TL;DR? We’ve got you covered:

Key security capabilities and strategies: Multi-factor authentication, cloud workload security, security analytics, encryption, identity and access management, cloud security gateways, microsegmentation, threat modeling, threat intelligence, and endpoint detection and response.

Keep things consistent: Develop a set of security policies and procedures to enforce on all of your clouds (and any on-prem software too, for that matter). While there will almost always be some kind of incompatibility, a benchmark or standardized security policy will reduce the risk of oversights.

Cloud agnostic software: Use security tools that can easily integrate with any cloud service, and that can scale with increased apps and workloads.

Go beyond your CSP’s tools: Your cloud providers have tools to keep their offerings safe, but protection of the data itself falls to you. Some vendors may be able to advise which capabilities you need within their infrastructure to keep your data safe.

Confidential computing: Data protection usually focuses on data at rest and in transit, but what about data in use? Protect data as it is being processed, and always know where the data is being used. Confidential computing will allow encrypted data to be processed in memory without exposing it to the rest of the system. This is a relatively new area, so consider keeping tabs on the Confidential Computing Consortium to stay in the loop.

Anticipate unforeseen changes: Planning for the unknown seems like an oxymoron – but in tech, it’s not. Things change constantly, and often in ways we don’t predict. Make sure your systems and environments can adapt to whatever the market throws at you.

Stay informed of new computing trends: For instance, Nick Ismail from Information Age highlights that serverless computing adoption is growing as it allows cloud instances to be scaled and patched instantly, and machine learning will be able to help servers identify patterns of malicious behavior and respond faster than human administrators can respond.

Looking ahead

Just like every market, cloud will continue to change as vendors make new alliances and focus on new capabilities. In 2020, Forrester predicts that hyperscale global public cloud leaders will form more alliances, while cloud management vendors will shift their focus to security – after a high-visibility data breach. Take steps to ensure that that breach isn’t yours by assessing the current and future state of your cloud strategy, and infusing security into everything you do.

Cover image by Michael Weidner on Unsplash.

Guide to the cloud Harness the power of the cloud with microservices, cloud-agnostic DevOps, and workflow portability. Learn more

Try all GitLab features - free for 30 days

GitLab is more than just source code management or CI/CD. It is a full software development lifecycle & DevOps tool in a single application.

Try GitLab for Free

Try GitLab risk-free for 30 days.

No credit card required. Have questions? Contact us.

Gitlab x icon svg