AWS has best of breed configurations and security services which firms should take advantage of. The following (13 areas), are the most important services to use when deploying into the AWS cloud.
Identity and Access Management, or IAM, must be used to administer access rights so root users can be assigned limited access but remain equipped to do the work required of their roles.
A root user is a user with full permission to view and change settings and configurations within an environment. Root user accounts are usually created to give access to the system for administrative functions, such as gathering information on billing and activity. It is often considered standard to grant these users unlimited access, but that broad level of control is not always needed.
With AWS IAM, users must be explicitly granted access to perform functions; no user is granted automatic access to anything. This allows companies to increase agility without incurring additional risk.
Removing root access from the system is a simple action that returns multiple security benefits. In addition to creating a more secure system overall, removing root and granting IAM access improves productivity of DevOps and product teams by letting them operate securely through the convenience and immediate management of their own AWS infrastructure security.
Businesses need more than the single layer of protection provided by usernames and passwords, which can be cracked, stolen or shared. Keep in mind that AWS IAM controls may provide access to not just the infrastructure but the applications installed and the data being used. By implementing MFA – a security measure that requires that a code be provided in addition to the password – you can assign roles, root accounts and IAM users securely.
MFA is commonly provided by way of physical or virtual devices separate from users’ username/password combinations. These devices generate random values to supplement the basic credentials, helping validate the identity of the person trying to access a cloud environment or some part of it. MFA devices can be separate, physical items or managed as smartphone apps. Aside from standard ID/password alphanumeric values, the biometrics required by some MFA continue to gain popularity, with many devices scanning various aspects that uniquely identify a user, such as retinas, fingerprints or other unique features.
Given the potential risk, an additional layer of security here makes good business sense. AWS is MFA-friendly, but it’s important that organizations use AWS guidelines to ensure compatibility of physical devices with their corresponding cloud infrastructure. When choosing an MFA product, consider both how it will integrate into your workflow and how to recover from its loss, as they do get lost and mobile devices do get replaced.
By limiting administrator access and aligning permission grants to the appropriate level of authority, you can minimize the risks of allowing too many users with administrator-level permissions.
When administrative users enter a company, they often introduce their own processes to security systems. As these administrators leave, it can often be difficult for new personnel to keep track of all the procedures in place, potentially leaving a company exposed to security risks.
By closely monitoring access levels and granting a limited number of user’s administrative access, you can optimize your security posture.
Using advanced technology such as IAM can help an organization eliminate the risks of security compromises.
Historically, security breaches have often occurred because users store their credentials in insecure locations, such as GitHub repositories. With advances in IAM, roles can now be granted temporary security tokens that will allow them to perform specific functions. With these tokens, roles can be used in combination with external identity providers, such as Security Assertion Markup Language, as wsell as to allow third parties to access resources. Since the access keys are not static, there is no need to store them. If a token is compromised, a new one can be issued, and if a token expires, the credential can be rotated.
With roles, users with lower levels of access can conduct tasks in Amazon Elastic Compute Cloud, or EC2, without needing to be granted excessive levels of access. This approach allows very specific access to AWS services and resources, reducing the possible attack surface area available to bad actors.
By employing IAM, you can limit the access of root users within your organization to the lowest level needed, minimizing the potential security risks to your company. Using the “least privilege” method to determine the lowest levels of access users need, you can eliminate the likelihood of compromising your company’s security position if a key is lost or stolen. In addition to the risks posed by users, network administrators must also keep in mind what applications and groups are granted permission to access their networks.
Limiting access based on the least privilege and IAM policies of your organization allow you to put layers of restrictions in place. These layers can act as the foundation of your organization, enabling you to eliminate risks posed to your network.
AWS CloudTrail is a tool that allows you to record API log information for security analysis, change tracking and compliance auditing. With it, you can create a trail of breadcrumbs to lead you back to the source of any changes made to your AWS environment.
AWS is an API-driven environment, so even if you use the AWS console, API calls are still being made behind the scenes on your behalf. When enabled, AWS CloudTrail logs details such as the time of the call; who made it, even if it is AWS or a third party; where it was made from, using the IP address; and whether it was successful.
When choosing whether to enable AWS CloudTrail, there are a few factors to consider:
If you use an S3 bucket to store CloudTrail logs, you will want to maintain records about all activity that touches or affects CloudTrail. With logging settings applied to the relevant S3 bucket, you’ll be able to track access requests as well as maintain a record of those who have access and the frequency with which they’re using it.
Since CloudTrail buckets contain sensitive information, these should be protected from unauthorized viewing. With S3 Server Access Logging enabled for your CloudTrail buckets, you can track any requests made to access the buckets. You can even limit who can alter or delete the access logs to prevent users from covering their tracks.
The corresponding information will include sensitive data and need to be protected. Like any other settings in AWS, this will provide more than a mechanism for insight into activity. It will also give you a way to modify, remove or grant access to logs to prevent users from making unauthorized changes to hide their identities or activities.
With Amazon EC2, systems running processes outside of it require keys, helping keep your system secure. Although roles re- move much of the need to manage keys, API keys should still be employed and regularly rotated. By rotating API keys regularly, you can control the amount of time for which a key is considered valid, limiting the impact to business if a compromise occurs.
The process of rotating keys does require some manual labor, but it takes only minimal effort from an administrator. With a system that is easily replicated, the process of alternating your keys every 90 days can be simplified by incorporating encrypted data snippets, such as Chef’s encrypted data bags.
Using roles for Amazon EC2 instances makes it easier for your resources to communicate securely and helps you reduce management burden by leveraging the AWS Security Token Service, or STS.
How often can you simultaneously become more secure and simplify management? These may sound like opposites, but they are what we strive for. Anytime you can make it easier for users to be more secure, you’re more likely to get adoption. Conversely, if you make security too complicated, it may reduce security in practice because people are inclined to take shortcuts.
At this point, we want to extend the roles for Amazon EC2 by using roles for your IAM users, making the instance more secure in a way that’s easier to maintain. We could go through each individual AWS account to create new users, generate passwords and restrict permissions – much like we had in the beginning – but that would be a lot of repetitive, manual steps. It’s better to leverage the users we’ve already created and secured, and just grant them access to the additional accounts.
If you have more than one AWS account, it’s worth the time to go through the steps outlined by AWS to get a good feel for what you can accomplish by making use of roles with your IAM users.
In addition to security, there are also many configuration best practices to follow. Per the Shared Responsibility Model, AWS is committed to the security of the AWS Cloud and is responsible for the foundation upon which applications on AWS are built.
However, as a user, you’re responsible for securing the product you create, the data you store on AWS, and how applications react to denial-of-service and distributed denial-of-service attacks.
DoS and DDoS attacks are attempts to make applications or websites inoperable by means of overwhelming them with traffic from multiple sources. DoS attacks are typically conducted by a single attacker, whereas DDoS attacks are characteristically directed by a group of collaborating parties.
The public nature of AWS makes any resource deployed on the system a target for DDoS attacks, so tools such as Elastic Load Balancing and Auto Scaling help prevent DDoS attacks from succeeding. Check out the “AWS Best Practices for DDoS Resiliency” white paper on how to mitigate DDoS attacks.
Although some organizations use application firewalls to deny DoS and DDoS attacks, Auto Scaling is a simpler, more cost- effective way to absorb such attacks and prevent service interruption. Most website traffic is directed by an elastic load balancer. As traffic to your website increases, the ELB undergoes corresponding scaling. Since ELBs only direct Transmission Control Protocol traffic, any attacks that use a protocol other than TCP will not reach your applications.
When TCP traffic enters your website, the data within that traffic must be analyzed. As the amount of data increases, you must be able to scale your ELB efficiently to avoid service interruption. To effectively analyze and accommodate your scaled ELB, you need a system with the capacity to scale with it automatically. Amazon EC2 has the capability to automatically employ its Auto Scaling function.
For example, let’s say you set a condition for Auto Scaling to launch two new Amazon EC2 application instances when the amount of network activity crosses a certain threshold. This trigger would already allow your site to scale based on normal, legitimate demand.
However, if abnormal attack traffic were to enter this site, Auto Scaling would also trigger a scale-up event, launching new Amazon EC2 instances to meet demand and process requests. This means your service remains operational during the attack and business can continue as normal.
Once the attack is over, Auto Scaling will automatically scale down the number of Amazon EC2 instances running simultaneously, if configured to do so. Although Auto Scaling means the price you pay for the increase in instance hours will rise, uninterrupted business operation is priceless. Auto Scaling acts an insurance policy that saves your company revenue, which justifies it in the end.
Auto Scaling is undoubtedly a best practice, but it may not be financially feasible for all organizations. For those without the resources to employ Auto Scaling, AWS Security Groups and Network Access Control Lists, or NACLs, provide a way to block hazardous website traffic that would otherwise trigger a need to scale.
Though they were not built to act as traditional firewalls, AWS Security Groups can provide the same functions as firewalls by allowing companies to simultaneously control network traffic accessibility and costs.
Much like a basic firewall, the principle purpose of an AWS Security Group is to allow traffic you want to admit to enter your network and move in the direction you would like it to flow. Depending on where the traffic is coming from, by employing an AWS Security Group, you can prevent malicious traffic from reaching your systems, stopping that traffic before attacks can commence.
Many companies deliver their products through web applications. For basic web applications, you may want to allow access by all types of traffic as not admitting this commerce could be detrimental to your business’s success.
In instances such as these, you should restrict your network’s response capabilities. To minimize the risk posed, you can effectively restrict your response to the two most common port numbers: Port 80 for unencrypted web traffic and Port 443 for encrypted traffic.
Most customers who access your application will never need administrative rights to your network. To continue to secure your network against compromise, you have two options: deploy infrastructure as code and never oversee a box manually; or only allow access from the origin IP and port your instance uses, and employ it when needed.
By only allowing access from the origin IP and port address through which you administer the instance, you can allow customers administrator access while maintaining your security posture. This method eliminates some of the convenience of automation, but you can ultimately reduce the attack surface area as well as the amount of time you’re exposed to potential attacks. This process can be scripted into your system for further convenience.
If you must allow unlimited access to your application, you will need to configure your network for both an NACL and an AWS Security Group.
Often, you’ll have to create special permissions for both an NACL and the AWS Security Group. By employing an AWS Security Group, you can limit the ports to which the world has access.
If you only need to allow specific network access to your application, you can limit access through the NACL, preventing anything you do not define from making it into your AWS Security Group.The NACL can also be thought of as the perimeter boundary, or your first checkpoint. With an NACL, you can configure an explicit deny rule to deny traffic.
AWS Security Groups are very focused on what you allow, denying everything that is not allowed. With an NACL configured, if you find you are receiving a DoS attack and the origin IPs are very specific, you can quickly and easily block it.
Although this process isn’t quite as seamless for a DDoS attack, if the attack is not evenly distributed, the set deny rules will likely still be able to hinder part of the attack. When employing either rule, familiarity with your logs and partnering with AWS Support will help you identify and repel attacks. Inbound rules are important, but the security of your network also relies on the outbound rules you’ve set in place. Configuring outbound allow rules, the process for which is no different from the process for configuring inbound rules, can be done with both AWS Security Groups and NACLs.
Once an AWS Security Group or NACL has been configured with implicit allow rules, anything not configured will be denied. Additionally, it’s important to keep in mind that AWS Security Groups are stateful while NACLs are not. As you configure outbound rules on your AWS Security Groups, they will not affect inbound sessions, but if you configure outbound rules on an NACL, you will need to allow the outbound traffic back to the origin IP to establish a session. It may be easier to configure your system so that AWS Security Groups use ports to access your network but NACLs are limited to your network. An AWS Security Group can be configured per application, whereas NACLs are designed to implement explicit rules for web applications.
For a two-tiered web application with web servers in one group and a database in another, security practices could be implemented by configuring inbound NACL rules to allow connections to web servers from around the world, while the database would only allow inbound connection from an established list of web servers. In this instance, web servers would only allow port 443 connections, while the database would only allow inbound 3306 connections for MySQL.
For outbound connections, you could remove both the web server and the database AWS Security Groups outbound rule to prevent them from initiating connections over the internet. The web servers would allow all outbound traffic out to ensure sessions could be established, and the database would limit outbound connections to the web server’s private subnet IP range.
Although IAM policies are built to provide security, organizations must take careful measures to ensure the stability of their platforms in the long run. As companies grow, they often add access control measures to their networks to keep up with the increasing demands placed on them.
As their networks expand, organizations often layer newer platforms on top of older systems, making it difficult to keep track of who or what is allowed access to their network.