All posts by Shiva Vaidyanathan

TLS inspection configuration for encrypted traffic and AWS Network Firewall

Post Syndicated from Shiva Vaidyanathan original https://aws.amazon.com/blogs/security/tls-inspection-configuration-for-encrypted-traffic-and-aws-network-firewall/

AWS Network Firewall is a managed service that provides a convenient way to deploy essential network protections for your virtual private clouds (VPCs). In this blog, we are going to cover how to leverage the TLS inspection configuration with AWS Network Firewall and perform Deep Packet Inspection for encrypted traffic. We shall also discuss key considerations and possible architectures.

Today, the majority of internet traffic is SSL/TLS encrypted to maintain privacy and secure communications between applications. Deep packet inspection (DPI) refers to the method of examining the full content of data packets as they traverse a network perimeter firewall. However, the lack of visibility into encrypted traffic presents a challenge to organizations that do not have the resources to decrypt and inspect network traffic. TLS encryption can hide malware, conceal data theft, or mask data leakage of sensitive information such as credit card numbers or passwords. Additionally, TLS decryption is compute-intensive and cryptographic standards are constantly evolving. Organizations that want to decrypt and inspect network traffic typically use a combination of hardware and software solutions from multiple vendors, which adds operational complexity and implementation challenges around capacity planning, scaling issues, and latency concerns. This forces some organizations to make adverse decisions to reduce the complexity of inspecting their network traffic such as blocking access to popular websites to mitigate performance problems.

There are multiple options you can use to perform DPI for encrypted traffic in your AWS environment, based on the use case. These include using AWS WAF or implementing third-party security appliances (next generation firewalls). The addition of new services like Gateway Load Balancer gives you more flexibility in designing your firewall architectures and the ability to perform DPI on AWS.

With this release, Network Firewall also becomes an option to support Deep Packet Inspection on encrypted payloads.

Considerations for deep packet inspection

The following are some key factors to consider when you enable TLS decryption functionality on Network Firewall.

DPI and performance. DPI is processor-intensive, because it not only looks into individual packets, but it also looks into traffic flows (a flow is a collection of related packets). This is combined with the fact that inspection needs to be done in real time with minimal impact to latency. Also, because many firewalls perform other advanced functions (for example, stateful packet inspection, NAT, virtual private network (VPN), and malware threat prevention), adding DPI increases the complexity of the entire system and impacts performance. However, because Network Firewall is an AWS managed service, the bandwidth performance of 100 gigabits per second (Gbps) per firewall endpoint is not impacted, even after you enable TLS inspection configuration. Single digit millisecond latency is expected at initial connection due to the TCP and TLS handshake before data can flow to the firewall. We recommend that you conduct your own testing for the rule sets to verify that the service meets your performance expectations.

DPI and encryption. Encryption has particularly been a challenge to DPI. Effective decisions can’t be made if the contents of the packets aren’t known. As more applications and websites use encryption, it is important that you implement the right TLS decryption technique. With Network Firewall, you can chose which traffic to decrypt by using your available certificates in AWS Certificate Manager (ACM). You can then apply the TLS configurations across the stateful rule groups, thereby authorizing Network Firewall to act as a go-between. For more information on how AWS Network Firewall handles privacy, please read the Network Firewall documentation.

AWS Network Firewall deployment architectures

There are three main architecture patterns for Network Firewall deployments. You can refer to the Deployment models for AWS Network Firewall blog post, which provides details on these, as well as key considerations. The three main models are as follows:

  • Distributed deployment model — Network Firewall is deployed into each individual VPC.
  • Centralized deployment model — Network Firewall is deployed into a centralized VPC for East-West (VPC-to-VPC) or North-South (inbound and outbound from internet, on-premises) traffic. We refer to this VPC as the inspection VPC throughout this blog post.
  • Combined deployment model — Network Firewall is deployed into a centralized inspection VPC for East-West (VPC-to-VPC) and a subset of North-South (on-premises, egress) traffic. Internet ingress is distributed to VPCs that require dedicated inbound access from the internet, and Network Firewall is deployed accordingly.

Each of these architectures is still valid for TLS inspection functionality. Today, AWS Network Firewall supports TLS inspection only for the ingress (inbound) traffic coming into the VPC.

In this section, we will highlight a deployment architecture with AWS Network Firewall and the process for deep packet inspection.

AWS Network Firewall – prior to TLS inspection configuration

Below figure 1 shows how Network Firewall performs inspection when the TLS inspection feature isn’t enabled. The workflow is as follows:

  1. The ingress traffic enters the VPC. Ingress routing enables the internet traffic to be inspected by AWS Network Firewall.
  2. The traffic from the firewall endpoint to the Network Firewall:
    1. Network Firewall inspects the packet first through a stateless engine. Network Firewall makes a drop/pass decision by applying the rules that are present in the stateless engine.
    2. If there is no match on the set of stateless rules present, the traffic is then forwarded to the stateful engine. Again, a drop/pass decision is made by applying the set of stateful rules.
  3. If the decision is to pass traffic, then the firewall endpoint present in the firewall subnet sends the traffic to the customer subnet through the routes present in the VPC subnet route table.

AWS Network Firewall — after TLS inspection configuration

After you enable the TLS inspection capability in Network Firewall, the traffic flow changes slightly, as shown in Figure 2. Because the ingress data you want to inspect is encrypted, it first needs to be decrypted before it is sent to the firewall stateful engine.

In Figure 2, you can see the ingress traffic flow, which has the following steps:

  1. The ingress traffic enters the VPC. Ingress routing enables the internet traffic to be inspected by AWS Network Firewall.
  2. The traffic from the firewall endpoint to the Network Firewall:
    1. Network Firewall inspects the packet first through a stateless engine. Network Firewall makes a drop/pass decision by applying the rules present in the stateless engine.
    2. If there is no match on the set of stateless rules present, the traffic is then forwarded to the stateful engine. However, before the traffic passes to the stateful engine, if there is no match and the traffic is in the scope of the TLS encryption configuration, the traffic is forwarded for the decrypt operation.
    3. After decryption, the traffic is then forwarded to the firewall stateful engine for inspection. Again, Network Firewall makes a drop/pass decision by applying the set of stateful rules.
  3. If the decision is to pass traffic, then the firewall endpoint present in the firewall subnet sends the traffic to the customer subnet through the routes present in the VPC subnet route table.

Note: Customers must trust this certificate for the TLS inspection configuration to function properly.

Implement TLS inspection in AWS Network Firewall

Let’s look at how to implement TLS inspection when you create a new network firewall in AWS Network Firewall. A TLS inspection configuration contains one or more references to a valid AWS Certificate Manager (ACM) SSL/TLS certificate that Network Firewall uses to decrypt ingress (inbound) traffic. Network Firewall supports a variety of certificate types supported in addition to wildcard certificates. You can optionally define a scope (5-tuple based) to decrypt traffic by source and destination IP or port. To follow this procedure, you must have at least one valid certificate type supported by Network Firewall in ACM that’s accessible by your AWS account.

To create a TLS inspection configuration (console)

  1. Sign in to the AWS Management Console and open the Amazon VPC console.
  2. In the navigation pane, under Network Firewall, choose TLS inspection configurations.
  3. Choose Create TLS inspection configuration.
    Figure 3: TLS inspection configuration for AWS Network Firewall

    Figure 3: TLS inspection configuration for AWS Network Firewall

  4. On the Associate SSL/TLS certificates page, in the search box, select the ACM certificate to use in the TLS inspection configuration, and then choose Add certificate. You can use as many as 10 certificates for a single configuration.
    Figure 4: SSL/TLS certificate as part of Network Firewall inspection configuration

    Figure 4: SSL/TLS certificate as part of Network Firewall inspection configuration

  5. Choose Next to go to the TLS inspection configuration’s Describe TLS inspection configuration page.
  6. For Name, enter a name to identify this TLS inspection configuration, and optionally enter a description for the TLS inspection configuration.
  7. Choose Next to go to the TLS inspection configuration’s Define scope page.
    Figure 5: Description for Network Firewall inspection configuration

    Figure 5: Description for Network Firewall inspection configuration

    Note that you can’t change the name after you create the TLS inspection configuration.

  8. In the Scope configuration pane, you can optionally define one or more 5-tuple scopes for the domains that you want Network Firewall to decrypt. Network Firewall uses the corresponding SSL/TLS certificates in your TLS inspection configuration to decrypt the SSL/TLS traffic that matches the scope criteria.
    Figure 6: Define scope for Network Firewall to decrypt

    Figure 6: Define scope for Network Firewall to decrypt

    • For Protocol, choose the protocol to decrypt and inspect for. Network Firewall currently supports only TCP.
    • For Source, choose the source IP addresses and ranges to decrypt and inspect for. You can inspect for either Custom IP addresses or Any IPv4 address. (IPv6 is currently not supported.)
    • For Source port, choose the source ports and source port ranges to decrypt and inspect for. You can inspect for either Custom port ranges or Any port.
    • For Destination, choose the destination IP addresses and ranges to decrypt and inspect for. You can inspect for either Custom IP addresses or Any IPv4 address.
    • For Destination port, choose the destination ports and destination port ranges to decrypt and inspect for. You can inspect for either Custom port ranges or Any port.
  9. After you’ve set the scope criteria, choose Add scope configuration, and then choose Next.
  10. On the next page, Select encryption options, determine whether you want to use the AWS managed key or customize encryption settings (advanced). Here we use the default key that AWS owns and manages on your behalf, choose Next.
    Figure 7: Select the encryption options

    Figure 7: Select the encryption options

  11. On the Add tags page, choose Next. Tags are optional but are recommended as a best practice. Tags help you organize and manage your AWS resources. For more information about tagging your resources, see Tagging AWS Network Firewall resources.
  12. On the Review and confirm page, check the TLS inspection configuration settings. Choose Create TLS inspection configuration. Your TLS inspection configuration is now ready for use.

    Figure 8: Validate the TLS inspection configuration

Update an existing network firewall with TLS inspection configuration

There are two methods that you can use modify an existing firewall configuration for TLS inspection, depending on your scenario.

Scenario 1: Add TLS inspection to an existing network firewall. In this scenario, you only need to consider the scope that TLS inspection applies to. After you have followed steps 1 through 12 outlined in the procedure in this post, and created the TLS inspection configuration, ingress (inbound) traffic will be decrypted and then sent to the stateful engine for inspection that uses your existing firewall policies.

Scenario 2: Modify an existing firewall with TLS inspection configured. In this scenario, where TLS configuration has already been added and you just need to modify the configuration, you can use the following steps. Note that you can’t change the name of a TLS inspection configuration after creation, but you can change other details.

To update a TLS inspection configuration

  1. Sign in to the AWS Management Console and open the Amazon VPC console.
  2. In the navigation pane, under Network Firewall, choose TLS inspection configurations.
  3. On the TLS inspection configuration page, select the name of the TLS inspection configuration you want to update.
  4. Make your desired changes to the configuration.
  5. Choose Save.

To understand more about how Network Firewall handles changes, including TLS inspection configuration, refer to Managing your firewall policy in AWS Network Firewall.

Conclusion

AWS Network Firewall lets you inspect traffic at scale in a variety of use cases. In this blog post, we looked into the recently launched TLS inspection configuration for ingress inspection architectures and discussed considerations for enabling this feature. We showed how you can enable and update the TLS inspection feature on Network Firewall. To learn more about the TLS inspection feature, check out the AWS Network Firewall Developer Guide. We hope this post is helpful and look forward to hearing about how you use the latest feature.

If you have feedback about this post, submit comments in the Comments section below. If you have questions about this post, contact AWS Support.

Want more AWS Security news? Follow us on Twitter.

Author

Shiva Vaidyanathan

Shiva is a Principal Cloud Infrastructure Architect at AWS. He provides technical guidance, design, and lead implementation projects to customers, to promote their success on AWS. He works towards making cloud networking secure and simple. Before joining AWS, he worked on several National Science Foundation (NSF)–funded research initiatives on how to perform secure computing in public cloud infrastructures. He holds an MS in Computer Science from Rutgers University and an MS in Electrical Engineering from New York University.

Brandon Carroll

Brandon Carroll

Brandon is a Senior Developer Advocate with AWS who is passionate about technology and sharing with the networking community. As a Developer Advocate at AWS, Brandon specializes in infrastructure security and helps customers in their journey to the cloud.

Deploy consistent DNS with AWS Service Catalog and AWS Control Tower customizations

Post Syndicated from Shiva Vaidyanathan original https://aws.amazon.com/blogs/architecture/deploy-consistent-dns-with-aws-service-catalog-and-aws-control-tower-customizations/

Many organizations need to connect their on-premises data centers, remote sites, and cloud resources. A hybrid connectivity approach connects these different environments. Customers with a hybrid connectivity network need additional infrastructure and configuration for private DNS resolution to work consistently across the network. It is a challenge to build this type of DNS infrastructure for a multi-account environment. However, there are several options available to address this problem with AWS. Automating DNS infrastructure using Route 53 Resolver endpoints covers how to use Resolver endpoints or private hosted zones to manage your DNS infrastructure.

This blog provides another perspective on how to manage DNS infrastructure with  Customizations for Control Tower and AWS Service Catalog. Service Catalog Portfolios and products use AWS CloudFormation to abstract the complexity and provide standardized deployments. The solution enables you to quickly deploy DNS infrastructure compliant with standard practices and baseline configuration.

Control Tower Customizations with Service Catalog solution overview

The solution uses the Customizations for Control Tower framework and AWS Service Catalog to provision the DNS resources across a multi-account setup. The Service Catalog Portfolio created by the solution consists of three Amazon Route 53 products: Outbound DNS product, Inbound DNS product, and Private DNS. Sharing this portfolio with the organization makes the products available to both existing and future accounts in your organization. Users who are given access to AWS Service Catalog can choose to provision these three Route 53 products in a self-service or a programmatic manner.

  1. Outbound DNS product. This solution creates inbound and outbound Route 53 resolver endpoints in a Networking Hub account. Deploying the solution creates a set of Route 53 resolver rules in the same account. These resolver rules are then shared with the organization via AWS Resource Access Manager (RAM). Amazon VPCs in spoke accounts are then associated with the shared resolver rules by the Service Catalog Outbound DNS product.
  2. Inbound DNS product. A private hosted zone is created in the Networking Hub account to provide on-premises resolution of Amazon VPC IP addresses. A DNS forwarder for the cloud namespace is required to be configured by the customer for the on-premises DNS servers. This must point to the IP addresses of the Route 53 Inbound Resolver endpoints. Appropriate resource records (such as a CNAME record to a spoke account resource like an Elastic Load Balancer or a private hosted zone) are added. Once this has been done, the spoke accounts can launch the Inbound DNS Service Catalog product. This activates an AWS Lambda function in the hub account to authorize the spoke VPC to be associated to the Hub account private hosted zone. This should permit a client from on-premises to resolve the IP address of resources in your VPCs in AWS.
  3. Private DNS product. For private hosted zones in the spoke accounts, the corresponding Service Catalog product enables each spoke account to deploy a private hosted zone. The DNS name is a subdomain of the parent domain for your organization. For example, if the parent domain is cloud.example.com, one of the spoke account domains could be called spoke3.cloud.example.com. The product uses the local VPC ID (spoke account) and the Network Hub VPC ID. It also uses the Region for the Network Hub VPC that is associated to this private hosted zone. You provide the ARN of the Amazon SNS topic from the Networking Hub account. This creates an association of the Hub VPC to the newly created private hosted zone, which allows the spoke account to notify the Networking Hub account.

The notification from the spoke account is performed via a custom resource that is a part of the private hosted zone product. Processing of the notification in the Networking Hub account to create the VPC association is performed by a Lambda function in the Networking Hub account. We also record each authorization-association within Amazon DynamoDB tables in the Networking Hub account. One table is mapping the account ID with private hosted zone IDs and domain name, and the second table is mapping hosted zone IDs with VPC IDs.

The following diagram (Figure 1) shows the solution architecture:

Figure 1. A Service Catalog based DNS architecture setup with Route 53 Outbound DNS product, Inbound DNS product, and Route 53 Private DNS product

Figure 1. A Service Catalog based DNS architecture setup with Route 53 Outbound DNS product, Inbound DNS product, and Route 53 Private DNS product

Prerequisites

Deployment steps

The deployment of this solution has two phases:

  1. Deploy the Route 53 package to the existing Customizations for Control Tower (CfCT) solution in the management account.
  2. Setup user access, and provision Route 53 products using AWS Service Catalog in spoke accounts.

All the code used in this solution can be found in the GitHub repository.

Phase 1: Deploy the Route 53 package to the existing Customizations for Control Tower solution in the management account

Log in to the AWS Management Console of the management account. Select the Region where you want to deploy the landing zone. Deploy the Customizations for Control Tower (CfCT) Solution.

1. Clone your CfCT AWS CodeCommit repository:

2. Create a directory in the root of your CfCT CodeCommit repo called route53. Create a subdirectory called templates and copy the Route53-DNS-Service-Catalog-Hub-Account.yml template and the Route53-DNS-Service-Catalog-Spoke-Account.yml under the templates folder.

3. Edit the parameters present in file Route53-DNS-Service-Catalog-Hub-Account.json with value appropriate to your environment.

4. Create a S3 bucket leveraging s3Bucket.yml template and customizations.

5. Upload the three product template files (OutboundDNSProduct.yml, InboundDNSProduct.yml, PrivateDNSProduct.yml) to the s3 bucket created in step 4.

6. Under the same route53 directory, create another sub-directory called parameters. Place the updated parameter json file from previous step under this folder.

7. Edit the manifest.yaml file in the root of your CfCT CodeCommit repository to include the Route 53 resource, manifest.yml is provided as a reference. Update the Region values in this example to the Region of your Control Tower. Also update the deployment target account name to the equivalent Networking Hub account within your AWS Organization.

8. Create and push a commit for the changes made to the CfCT solution to your CodeCommit repository.

9. Finally, navigate to AWS CodePipeline in the AWS Management Console to monitor the progress. Validate the deployment of resources via CloudFormation StackSets is complete to the target Networking Hub account.

Phase 2: Setup user access, and provision Route 53 products using AWS Service Catalog in spoke accounts

In this section, we walk through how users can vend products from the shared AWS Service Catalog Portfolio using a self-service model. The following steps will walk you through setting up user access and provision products:

1. Sign in to AWS Management Console of the spoke account in which you want to deploy the Route 53 product.

2. Navigate to the AWS Service Catalog service, and choose Portfolios.

3. On the Imported tab, choose your portfolio as shown in Figure 2.

Figure 2. Imported DNS portfolio (spoke account)

Figure 2. Imported DNS portfolio (spoke account)

4. Choose the Groups, roles, and users pane and add the IAM role, user, or group that you want to use to launch the product.

5. In the left navigation pane, choose Products as shown in Figure 3.

6. On the Products page, choose either of the three products, and then choose Launch Product.

Figure 3. DNS portfolio products (Inbound DNS, Outbound DNS, and Private DNS products)

Figure 3. DNS portfolio products (Inbound DNS, Outbound DNS, and Private DNS products)

7. On the Launch Product page, enter a name for your provisioned product, and provide the product parameters:

  • Outbound DNS product:
    • ChildDomainNameResolverRuleId: Rule ID for the Shared Route 53 Resolver rule for child domains.
    • OnPremDomainResolverRuleID: Rule ID for the Shared Route 53 Resolver rule for on-premises DNS domain.
    • LocalVPCID: Enter the VPC ID, which the Route 53 Resolver rules are to be associated with (for example: vpc-12345).
  • Inbound DNS product:
    • NetworkingHubPrivateHostedZoneDomain: Domain of the private hosted zone in the hub account.
    • LocalVPCID: Enter the ID of the VPC from the account and Region where you are provisioning this product (for example: vpc-12345).
    • SNSAuthorizationTopicArn: Enter ARN of the SNS topic belonging to the Networking Hub account.
  • Private DNS product:
    • DomainName: the FQDN for the private hosted zone (for example: account1.parent.internal.com).
    • LocalVPCId: Enter the ID of the VPC from the account and Region where you are provisioning this product.
    • AdditionalVPCIds: Enter the ID of the VPC from the Network Hub account that you want to associate to your private hosted zone.
    • AdditionalAccountIds: Provide the account IDs of the VPCs mentioned in AdditionalVPCIds.
    • NetworkingHubAccountId: Account ID of the Networking Hub account
    • SNSAssociationTopicArn: Enter ARN of the SNS topic belonging to the Networking Hub account.

8. Select Next and Launch Product.

Validation of Control Tower Customizations with Service Catalog solution

For the Outbound DNS product:

  • Validate the successful DNS infrastructure provisioning. To do this, navigate to Route 53 service in the AWS Management Console. Under the Rules section, select the rule you provided when provisioning the product.
  • Under that Rule, confirm that spoke VPC is associated to this rule.
  • For further validation, launch an Amazon EC2 instance in one of the spoke accounts.  Resolve the DNS name of a record present in the on-premises DNS domain using the dig utility.

For the Inbound DNS product:

  • In the Networking Hub account, navigate to the Route 53 service in the AWS Management Console. Select the private hosted zone created here for inbound access from on-premises. Verify the presence of resource records and the VPCs to ensure spoke account VPCs are associated.
  • For further validation, from a client on-premises, resolve the DNS name of one of your AWS specific domains, using the dig utility, for example.

For the Route 53 private hosted zone (Private DNS) product:

  • Navigate to the hosted zone in the Route 53 AWS Management Console.
  • Expand the details of this hosted zone. You should see the VPCs (VPC IDs that were provided as inputs) associated during product provisioning.
  • For further validation, create a DNS A record in the Route 53 private hosted zone of one of the spoke accounts.
  • Spin up an EC2 instance in the VPC of another spoke account.
  • Resolve the DNS name of the record created in the previous step using the dig utility.
  • Additionally, the details of each VPC and private hosted zone association is maintained within DynamoDB tables in the Networking Hub account

Cleanup steps

All the resources deployed through CloudFormation templates should be deleted after successful testing and validation to avoid any unwanted costs.

  • Remove the changes made to the CfCT repo to remove the references to the Route 53 folder in the manifest.yaml and the route53 folder. Then commit and push the changes to prevent future re-deployment.
  • Go to the CloudFormation console, identify the stacks appropriately, and delete them.
  • In spoke accounts, you can shut down the provisioned AWS Service Catalog product(s), which would terminate the corresponding CloudFormation stacks on your behalf.

Note: In a multi account setup, you must navigate through account boundaries and follow the previous steps where products were deployed.

Conclusion

In this post, we showed you how to create a portfolio using AWS Service Catalog. It contains a Route 53 Outbound DNS product, an Inbound DNS product, and a Private DNS product. We described how you can share this portfolio with your AWS Organization. Using this solution, you can provision Route 53 infrastructure in a programmatic, repeatable manner to standardize your DNS infrastructure.

We hope that you’ve found this post informative and we look forward to hearing how you use this feature!

Hands-on walkthrough of the AWS Network Firewall flexible rules engine – Part 2

Post Syndicated from Shiva Vaidyanathan original https://aws.amazon.com/blogs/security/hands-on-walkthrough-of-the-aws-network-firewall-flexible-rules-engine-part-2/

This blog post is Part 2 of Hands-on walkthrough of the AWS Network Firewall flexible rules engine – Part 1. To recap, AWS Network Firewall is a managed service that offers a flexible rules engine that gives you the ability to write firewall rules for granular policy enforcement. In Part 1, we shared how to write a rule and how the rule engine processes rules differently depending on whether you are performing stateless or stateful inspection using the action order method.

In this blog, we focus on how stateful rules are evaluated using a recently added feature—the strict rule order method. This feature gives you the ability to set one or more default actions. We demonstrate how you can use this feature to create or update your rule groups and share scenarios where this feature can be useful.

In addition, after reading this post, you’ll be able to deploy an automated serverless solution that retrieves the latest Suricata-specific rules from the community, such as from Proofpoint’s Emerging Threats OPEN ruleset. By deploying such solutions into your Amazon Web Services (AWS) environment, you can seamlessly enhance your overall security posture as the solutions fetch the latest set of intrusion detection system (IDS) rules from Proofpoint (formerly Emerging Threats) and optionally using them as intrusion prevention system (IPS) thereby keeping the rule groups updated on your Network Firewall. You can select the refresh interval to update these rulesets—the default refresh interval is 6 hours. You can also convert the set of rule groups to intrusion prevention system (IPS) mode. Finally, you have granular visibility of the various categories of rules for your Network Firewall on the AWS Management Console.

How does Network Firewall evaluate your stateful rule group?

There are two ways that Network Firewall can evaluate your stateful rule groups: the action ordering method or the strict ordering method. The settings of your rule groups must match the settings of the firewall policy that they belong to.

With the action order evaluation method for stateless inspection, all individual packets in a flow are evaluated against each rule in the policy. The rules are processed in order based on the priority assigned to them with lowest numbered rules evaluated first. For stateful inspection using the action order evaluation method, the rule engine evaluates based on the order of their action setting with pass rules processed first, then drop, then alert. The engine stops processing rules when it finds a match. The firewall also takes into consideration the order that the rules appear in the rule group, and the priority assigned to the rule, if any. Part 1 provides more details on action order evaluation.

If your firewall policy is set up to use strict ordering, Network Firewall now allows you the option to manually set a strict rule group order for stateful rule groups. Using this optional setting, the rule groups are evaluated in order of priority, starting from the lowest numbered rule, and the rules in each rule group are processed in the order in which they’re defined. You can also select which of the default actions—drop all, drop established, alert all, or alert established—Network Firewall will take when following strict rule ordering.

A customer scenario where strict rule order is beneficial

Configuring rule groups by action order is appropriate for IDS use cases, but can be an obstacle for use cases where you deploy firewalls that follow security best practice, which is to allow only what’s required and deny everything else (default deny). You can’t achieve this best practice by using the default action order behavior. However, with strict order functionality, you can create a firewall policy that allows prioritization of stateful rules, or that can run 5-tuple and Suricata rules simultaneously. Strict rule order allows you to have a block of fine-grain rules with specific actions at the beginning followed by a coarse set of rules with specific actions and finally a default drop action. An example is shown in Figure 1 that follows.

Figure 1: An example snippet of a Network Firewall firewall policy with strict rule order

Figure 1: An example snippet of a Network Firewall firewall policy with strict rule order

Figure 1 shows that there are two different default drop actions that you can choose:
drop established and
drop all. If you choose
drop established, Network Firewall drops only the packets that are in established connections. This allows the layer 3 and 4 connection establishment packets that are needed for the upper-layer connections to be established, while dropping the packets for connections that are already established. This allows application-layer
pass rules to be written in a default-deny setup without the need to write additional rules to allow the lower-layer handshaking parts of the underlying protocols.

The drop all action drops all packets. In this scenario, you need additional rules to explicitly allow lower-level handshakes for protocols to succeed. Evaluation order for stateful rule groups provides details of how Network Firewall evaluates the different actions. In order to set the additional environment variables that are shown in the snippet, follow the instructions outlined in Examples of stateful rules for Network Firewall and the Suricata rule variables.

An example walkthrough to set up a Network Firewall policy with a stateful rule group with strict rule order and default drop action

In this section, you’ll start by creating a firewall policy with strict rule order. From there, you’ll build on it by adding a stateful rule group with strict rule order and modifying the priority order of the rules within a stateful rule group.

Step 1: Create a firewall policy with strict rule order

You can configure the default actions on policies using strict rule order, which is a property that can only be set at creation time as described below.

  1. Log in to the console and select the AWS Region where you have Network Firewall.
  2. Select VPC service on the search bar.
  3. On the left pane, under the Network Firewall section, select Firewall policies.
  4. Choose Create Firewall policy. In Describe firewall policy, enter an appropriate name and (optional) description. Choose Next.
  5. In the Add rule groups section.
    1. Select the Stateless default actions:
      1. Under Choose how to treat fragmented packets choose one of the options.
      2. Choose one of the actions for stateless default actions.
    2. Under Stateful rule order and default action
      1. Under Rule order choose Strict.
      2. Under Default actions choose the default actions for strict rule order. You can select one drop action and one or both of the alert actions from the list.
  6. Next, add an optional tag (for example, for Key enter Name, and for Value enter Firewall-Policy-Non-Production). Review and choose Create to create the firewall policy.

Step 2: Create a stateful rule group with strict rule order

  1. Log in to the console and select the AWS Region where you have Network Firewall.
  2. Select VPC service on the search bar.
  3. On the left pane, under the Network Firewall section, select Network Firewall rule groups.
  4. In the center pane, select Create Network Firewall rule group on the top right.
    1. In the rule group type, select Stateful rule group.
    2. Enter a name, description, and capacity.
    3. In the stateful rule group options select either 5-tuple or Suricata compatible IPS rules. These allow rule order to be strict.
    4. In the Stateful rule order, choose Strict.
    5. In the Add rule section, add the stateful rules that you require. Detailed instructions on creating a rule can be found at Creating a stateful rule group.
    6. Finally, Select Create stateful rule group.

Step 3: Add the stateful rule group with strict rule order to a Network Firewall policy

  1. Log in to the console and select the AWS Region where you have Network Firewall.
  2. Select VPC service on the search bar.
  3. On the left pane, under the Network Firewall section, select Firewall policies.
  4. Chose the network firewall policy you created in step 1.
  5. In the center pane, in the Stateful rule groups section, select Add rule group.
  6. Select the stateful rule group you created in step 2. Next, choose Add stateful rule group. This is explained in detail in Updating a firewall policy.

Step 4: Modify the priority of existing rules in a stateful rule group

  1. Log in to the console and select the AWS Region where you have Network Firewall.
  2. Select VPC service on the search bar.
  3. On the left pane, under the Network Firewall section, choose Network Firewall rule groups.
  4. Select the rule group that you want to edit the priority of the rules.
  5. Select the Edit rules tab. Select the rule you want to change the priority of and select the Move up and Move down buttons to reorder the rule. This is shown in Figure 2.

 

Figure 2: Modify the order of the rules within a stateful rule groups

Figure 2: Modify the order of the rules within a stateful rule groups

Note:

  • Rule order can be set to strict order only when network firewall policies or rule groups are created. The rule order can’t be changed to strict order evaluation on existing objects.
  • You can only associate strict-order rule groups with strict-order policies, and default-order rule groups with default-order policies. If you try to associate an incompatible rule group, you will get a validation exception.
  • Today, creating domain list-type rule groups using strict order isn’t supported. So, you won’t be able to associate domain lists with strict order policies. However, 5-tuple and Suricata compatible rules are supported.

Automated serverless solution to retrieve Suricata rules

To help simplify and maintain your more advanced Network Firewall rules, let’s look at an automated serverless solution. This solution uses an Amazon CloudWatch Events rule that’s run on a schedule. The rule invokes an AWS Lambda function that fetches the latest Suricata rules from Proofpoint’s Emerging Threats OPEN ruleset and extracts them to an Amazon Simple Storage Service (Amazon S3) bucket. Once the files lands in the S3 bucket another Lambda function is invoked that parses the Suricata rules and creates rule groups that are compatible with Network Firewall. This is shown in Figure 3 that follows. This solution was developed as an AWS Serverless Application Model (AWS SAM) package to make it less complicated to deploy. AWS SAM is an open-source framework that you can use to build serverless applications on AWS. The deployment instructions for this solution can be found in this code repository on GitHub. 

Figure 3: Network Firewall Suricata rule ingestion workflow

Figure 3: Network Firewall Suricata rule ingestion workflow

Multiple rule groups are created based on the Suricata IDS categories. This solution enables you to selectively change certain rule groups to IPS mode as required by your use case. It achieves this by modifying the default action from alert to drop in the ruleset. The modified stateful rule group can be associated to the active Network Firewall firewall policy. The quota for rule groups might need to be increased to incorporate all categories from Proofpoint’s Emerging Threats OPEN ruleset to meet your security requirements. An example screenshot of various IPS categories of rule groups created by the solution is shown in Figure 4. Setting up rule groups by categories is the preferred way to define an IPS rule, because the underlying signatures have already been grouped and maintained by Proofpoint.   

Figure 4: Rule groups created by the solution based on Suricata IPS categories

Figure 4: Rule groups created by the solution based on Suricata IPS categories

The solution provides a way to use logs in CloudWatch to troubleshoot the Suricata rulesets that weren’t successfully transformed into Network Firewall rule groups.
The final rulesets and discarded rules are stored in an S3 bucket for further analysis. This is shown in Figure 5. 

Figure 5: Amazon S3 folder structure for storing final applied and discarded rulesets

Figure 5: Amazon S3 folder structure for storing final applied and discarded rulesets

Conclusion

AWS Network Firewall lets you inspect traffic at scale in a variety of use cases. AWS handles the heavy lifting of deploying the resources, patch management, and ensuring performance at scale so that your security teams can focus less on operational burdens and more on strategic initiatives. In this post, we covered a sample Network Firewall configuration with strict rule order and default drop. We showed you how the rule engine evaluates stateful rule groups with strict rule order and default drop. We then provided an automated serverless solution from Proofpoint’s Emerging Threats OPEN ruleset that can aid you in establishing a baseline for your rule groups. We hope this post is helpful and we look forward to hearing about how you use these latest features that are being added to Network Firewall.

Author

Shiva Vaidyanathan

Shiva is a senior cloud infrastructure architect at AWS. He provides technical guidance, and designs and leads implementation projects for customers to ensure their success on AWS. He works towards making cloud networking simpler for everyone. Prior to joining AWS, he worked on several NSF-funded research initiatives on how to perform secure computing in public cloud infrastructures. He holds a MS in Computer Science from Rutgers University and a MS in Electrical Engineering from New York University.

Author

Lakshmikanth Pandre

Lakshmikanth is a senior technical consultant with an AWS Professional Services team based out of Dallas, Texas. With more than 20 years of industry experience, he works as a trusted advisor with a broad range of customers across different industries and segments, helping the customers on their cloud journey. He focuses on design and implementation, and he consults on devops strategies, infrastructure automation, and security for AWS customers.

Author

Brian Lazear

Brian is head of product management for AWS Network Firewall and Firewall Manager services. He has over 15 years of experience helping enterprise customers build secure applications in the cloud. In AWS, his focus is on network security, firewalls, NDR/EDR, monitoring, and traffic-mirroring services.