Tuesday, October 8, 2024

Radiator Software and Altice Labs announce partnership

Altice Labs, a technology company that is at the forefront of global innovative solutions in telecommunications, networks, and digital services, alongside Radiator Software, a Finnish company which provides AAA (RADIUS/Diameter Authentication, Authorization and Accounting) software products and services for Service Providers and Enterprises, announced a partnership enabling both organizations to jointly deliver end-to-end solutions to Service Providers and Enterprises.

This partnership will allow Altice Labs and Radiator Software to combine efforts to improve efficiency by eliminating barriers and accelerating delivery, thereby enhancing the value of products and solutions for both organizations. One of the key use cases that Radiator Software and Altice Labs can provide together includes WiFi offload and Voice-over-WiFi solutions, among others.

Jaakko Stenhäll, Director of Business Development at Radiator Software, highlighted that, “with Altice Labs, we are able to complement our Radiator AAA offering with excellent technical knowledge on different customer needs and also world-class support on various markets - bringing great value to our mutual customers”.

Tiago Pereira, Director of Global Business Development at Altice Labs, commented, “we are looking forward to partnering with Radiator Software, bringing our extensive experience, knowledge and technical expertise on network and service management and control. This is an area where Altice Labs has been present for more than 20 years, with its own products and solutions”.

For Cleverson Novo, Managing Director of Open Labs - Altice Labs branch in Brazil, “this area has gained huge importance in the Latin America market, with Service Providers leveraging their Wi-Fi networks as a complement to traditional cellular networks“.

Together, both companies are paving the way to the future, exploring new marketing opportunities and business growth.

Wednesday, September 25, 2024

Meet Radiator at WGC EMEA & Network X in Paris!

We are delighted to announce that Radiator Software will be attending the two top connectivity events of the season: WGC EMEA and Network X, co-located in Paris on 7 – 10 October 2024. We are looking forward to meeting our current and prospective partners and customers in Paris.

Wireless Global Congress Americas

WGC is hosted by the Wireless Broadband Alliance and gathers together Wi-Fi industry leaders and experts from all around the world. As usual, the event is divided into two parts: WBA Members-Only Sessions on 7 and 8 October at the Hôtel Mercure Paris Porte de Versailles Expo, followed by the WGC EMEA Open Congress on 9 and 10 October in Porte de Versailles conference centre.

https://www.wirelessglobalcongress.com/emea-2024/

Network X

Network X event runs through 8 - 10 October and brings together Broadband World Forum, 5G World and Telco Cloud. For service providers of all kinds, Radiator provides a flexible AAA solution for fixed broadband, wireless, and WiFi offloading including VoWiFi.

For more information about the Network X event, please see the official website: https://networkxevent.com/

Meet with Radiator team

You can find the Radiator team at the WBA Members-Only Sessions and in the WGC Open Congress throughout the event. For insights on Wi-Fi authentication, Wi-Fi roaming and OpenRoaming, we extend an invitation to all WGC EMEA and Network X attendees to meet with the Radiator team and managing director Karri Huhtanen.

To book a meeting or simply ask a question, please leave a message and we will get back to you. See you in Paris!

Tuesday, September 17, 2024

Chargeable User Identity - Billing and analytics with privacy

As mobile and wireless networks have evolved, the ability for users to move between different networks while maintaining service, known as roaming, has become essential. WiFi roaming, while convenient for users, introduces several complexities for service providers, particularly in managing billing and user identity securely across network boundaries.

The Chargeable User Identity (CUI) parameter was introduced to address these challenges. While the specification RFC 4372 for CUI has existed for quite awhile, the implementations are now popularising as commercial Wi-Fi is becoming more sought after.

Chargeable User Identity uses and benefits

Chargeable user identity is a parameter used mostly by service providers to identify users for accounting in roaming networks, while ensuring their privacy is not compromised with trackable credentials. The CUI allows service providers to charge users based on their usage, even when users roam across different networks. It is primarily intended for billing purposes, but also provides other benefits to both public and commercial networks.

The main benefit of using Chargeable User Identity parameter is that it solves the business problem of anonymity in commercial networks, while not making any compromises in privacy and security. It provides a robust mechanism for calculating usage, which can be used not only for billing but also analytics purposes. For example, with CUI, roaming network providers can track whether their 100 sessions come from 10 users with 10 sessions each or by 2 users with 50 sessions each. This allows for more accurate analytics, but does not allow the networks to identify the users. This is possible in deployments where the CUI is the same across all of the user’s devices.

The use of Chargeable User Identity also allows public institutions to restrict and ban roaming users who violate their terms. Previously, when administrators decided to take action against users who violate their visiting terms, the user can simply log on with another device. With a CUI parameter that is mutual across user’s all devices, this is not possible.

Chargeable User Identity deployment

Chargeable User Identity is transmitted in RADIUS packets using dedicated RADIUS attribute 89: Chargeable-User-Identity. The implementation is specified in RFC 4372.

Upon sending the authentication Access-Request to the home organisation for a roaming user’s authentication, the visiting organisation should add the Chargeable-User-Identity parameter into the request with a null value. This signals the home organisation that a CUI is requested. The home organisation check’s for an existing valid CUI and sends either a new or existing valid CUI included in the Access-Accept.

The Chargeable-User-Identity parameter will remain the same for the duration of the roaming user’s session and is included in the accounting packets and responses.

Want to know more?

Are you looking to deploy Wi-Fi offloading or other Wi-Fi roaming functionality for your customers or members of your organisation? Or are you setting up a commercial Wi-Fi infrastructure to provide roaming services for operators? For both cases, Radiator AAA is the product for you.

Radiator AAA provides functionality for Wi-Fi roaming host organisations, with dozens of completed deployments for the biggest Wi-Fi roaming networks (eduroam, govroam, OpenRoaming). Combined with the Radiator SIM Pack, Radiator provides seamless authentication for Wi-Fi offloading, roaming between mobile networks and Wi-Fi. Both products and use cases include Chargeable User Identity function for Radiator.

For more information about CUI deployments, please contact our sales team at sales@radiatorsoftware.com

Wednesday, August 28, 2024

Why 5G drives Wi-Fi offloading for operators?

Latest reports from the industry indicate that globally 5G subscriptions are closing in on 20% of all mobile subscriptions. Mobile operators are deploying 5G networks at an increasing rate and 5G is at an early stage of its life cycle. Global data usage increases year by year and mobile operators’ cellular networks are hard-pressed to withstand all the traffic, requiring investment in more infrastructure.

5G provides higher data rates and other benefits compared to the previous generations, but at the cost of lower signal range. Strong 5G signal for proper coverage requires operators investment in small cell networks and even with one customers might struggle with in-door service quality. At the same time, there are commercial and other WiFi network infrastructure already in place for many of the areas where operators struggle with in-door coverage.

In this blog post, we’ll take a look at underlying reasons driving the demand for Wi-Fi offloading, and how Radiator SIM Pack solution provides the seamless authentication to enable it. In short, with the performance and reliability that Wi-Fi 6 and 7 have brought to the table, it is easy to see the economical and environmental benefits Wi-Fi offloading brings to those who adopt it, all while improving end users’ coverage and quality of service indoors without need for massive infrastructure investments.

Solve coverage issues with existing infrastructure

Even before 5G, operators were struggling with in-door coverage of cellular networks. Upgrading to 5G will not improve the coverage, but rather do the exact opposite. 5G utilises higher signal frequency millimetre waves, which are unable to penetrate obstructions and have short range. This is becoming a key concern when designing 5G networks in congested areas. However, building a network of small cells to reach proper coverage in-doors is fast becoming a challenge for operators in urban areas.

Wi-Fi offloading can significantly enhance coverage and quality of service for network operators by leveraging the ubiquitous presence of Wi-Fi networks to alleviate congestion on cellular networks. By directing data traffic from overloaded cellular networks to available Wi-Fi connections, operators can effectively extend coverage into areas with weak cellular signal and manage high-demand scenarios, such as large public venues. Wi-Fi offloading not only optimises the utilisation of network resources but additionally offloading traffic to Wi-Fi can help operators reduce network congestion and associated operational costs, making it a win-win solution for both service providers and their customers.

As an added benefit for the end user, According to Wireless Broadband Association, smartphones and IoT devices using Wi-Fi 6 have an up to 67% lower power consumption compared to their respective cellular networks. This energy efficiency will be further enhanced with Wi-Fi 7. This does not directly affect the operator, but enhances the end users’ service quality.

Wi-Fi has gotten better. Much better.

This point is not specific to 5G, but rather for all operators who have previously considered Wi-Fi offloading and found Wi-Fi to have high latency, unreliable connections and low data rate, not matching the standard operators want for their networks. This may have been the case once, but not anymore.

Wi-Fi 6, 6E and especially Wi-Fi 7 have brought down latency (below 5ms on most estimations), increased data rate tremendously (up to 46 Gbps) and made connections much more reliable. And as discussed previously, Wi-Fi networks are built with coverage in mind, bringing offloading users optimal quality of service even in areas with dozens of devices online.

Another concern our MNO customers have expressed from time to time is that Wi-Fi security is not up to par with mobile networks. Today’s Wi-Fi offloading solutions use EAP-AKA and EAP-AKA’ authentication, which provides vast improvements to older protocols. IMSI Privacy and standards in MAC address randomisation should be in place in a modern Wi-Fi offloading solution, with which end users details remain private and can not be snooped. Security and privacy concerns are a thing of the past for operators who want to adopt Wi-Fi offloading, so long as the operator chooses a solution that provides these features.

5G and Wi-Fi 7: Better together

We’ve seen many industry blogs and articles discuss the differences of 5G and Wi-Fi, often comparing them as rivals and recommending customers choose one or the other based on their needs. This need not be the case. As leading operators have demonstrated, these networks are not at odds, but rather better utilised together.

Wi-Fi offloading can solve operators’ problem of coverage in congested areas, particularly in-doors and in large venues with high volumes of data traffic. These areas often have Wi-Fi infrastructure in place which operators can utilise, lessening the need for investment in small cell networks.

As an improvement to previous mobile network generations, Wi-Fi offloading has been clearly specified as part of 5G architecture in 3GPP standards (3GPP TS 33.501; Annex S). When viewed as part of the architecture with dedicated authentication interface, rather than a case-by-case solution, Wi-Fi offloading is becoming a more integral part of operators’ connectivity stack.

Looking to deploy Wi-Fi offloading in your network?

Are you looking to adopt Wi-Fi offloading to your cellular network? Radiator SIM Pack is the product for you! Radiator SIM Pack provides seamless authentication for mobile users roaming between cellular and wireless networks.

Radiator SIM Pack provides SIM-based authentication (EAP-SIM, EAP-AKA, EAP-AKA’) with IMSI Privacy and a variety of different integration options for Diameter interfaces and for logging. Often combined with the Radiator Policy and Charging Pack for OCS billing integration, these products provide a one stop shop for operators looking to adopt seamless Wi-Fi offloading for their mobile customers.

If you wish to learn more about our Wi-Fi offloading deployments, please do not hesitate to contact sales@radiatorsoftware.com

Tuesday, July 9, 2024

Radiator 4.29 released!

We are pleased to announce the release of Radiator version 4.29. The latest release includes a major Radius protocol security fix, and the usual usability and interoperability improvements and bug fixes.

New usability improvements

  • Tested and supported for Ubuntu 24.04
  • AuthBy LDAP2 improvements
  • CEF and JSON logging fixes
  • Updates to address BlastRADIUS protocol vulnerability

    Radiator is actively engaged with IETF’s radext working group and we have been working under embargo to implement the fixes based on the work done in the group.

  • Add a new flag parameter LimitProxyState to Client clauses. This parameter allows dropping those requests from non-proxy clients that contain Proxy-State but do not contain Message-Authenticator. Ensure that ServeRADSEC drops requests with bad Message-Authenticator instead of just logging them. The upcoming Radius transport update by IETF's radext working group will remove the redundant signatures but keep them for the current transport profile. LimitProxyState addresses CVE-2024-3596.
  • Update RADIUS Message-Authenticator attribute handling. Message-Authenticator is always added as the first attribute in Radius messages. Message-Authenticator is now added automatically to replies to Access-Request messages and to Access-Request messages when they are proxied.
  • RequireMessageAuthenticator is now available for AuthBy RADIUS and its subclasses. It can be set for all hosts in an AuthBy or host-by-host basis. This parameter requires a valid Message-Authenticator in proxy replies.
  • A new configuration flag -no_message_authenticator is available in radpwtst to skip Message-Authenticator in Access-Requests.
  • Please see the security notice for more information on CVE-2024-3596 and our security recommendations.

    New attributes ensuring interoperability

  • Vendor specific attributes updated in the Radiator dictionary for Arista, Dell, ELTEK, Force10, Mojo, and Teldat.
  • More detailed changes can be found in the revision history. Radiator packages are available to download for current licensees from the downloads page and the Radiator repository.

    Would you like to know more?

    As always, you can contact Radiator team at info(a)radiatorsoftware.com - we are happy to learn more about your use case and assist you!

    Security Notice: BlastRADIUS protocol vulnerability (CVE-2024-3596) fixed in Radiator v4.29

    In February 2024 University of California San Diego researchers and their partners reported a vulnerability discovered in the RADIUS protocol to the CERT and IETF. The RADIUS protocol vulnerability was later named BlastRADIUS. The vulnerability allows the attacker to alter RADIUS messages between the RADIUS server and client so that for example a rejected authentication can be turned into accepted authentication. Utilising the vulnerability requires that the attacker is able replace the original requests and replies between RADIUS client and server with the attacker’s content. Only the RADIUS messages, which do not contain message authenticator, or where the RADIUS client or server is not verifying message authenticator properly, are vulnerable.

    Radiator Software has been working since February together with the researchers and other RADIUS server implementers to implement the identified fixes for RADIUS protocol in Radiator. This work and fixes have been under embargo until July 9th 2024 12:00am UTC. The fixes for RADIUS protocol have been implemented in the Radiator v4.29 released now after the embargo has ended.

    Affected Radiator versions

    Since this is a recently discovered RADIUS protocol vulnerability, all Radiator versions up until version 4.29 are affected.

    Affected Radiator configurations and deployments

    The RADIUS server deployments, which send and receive RADIUS requests without message authenticator or RadSec (RADIUS over TLS) over untrusted networks, are the most vulnerable for abuse. Enterprise Wi-Fi (EAP) authentication is using message authenticator in all messages by default so in those networks, the effect of the vulnerability depends if the RADIUS client implementation (e.g. in a Wi-Fi controller) is verifying the message authenticator properly. In addition to the RADIUS server updates, also RADIUS clients and client software may need to be updated.

    As Radiator implements the RADIUS protocol all Radiator configurations and deployments using RADIUS protocol without requiring Message Authenticator or RadSec (RADIUS over TLS) are affected by this vulnerability.

    Mitigation

    Radiator Software strongly recommends upgrading Radiator to the latest version to get all the improvements and fixes to the vulnerability. In addition to upgrading Radiator, the Message-Authenticator requirement instructions should also be followed.

    Configuring Message-Authenticator requirement

    Requiring Message-Authenticator from RADIUS clients

    Radiator version 4.10 (2012) or newer can be configured to require a Message-Authenticator from a RADIUS client by adding RequireMessageAuthenticator to the Client configuration. For example:

    <Client 192.0.2.42>
         Identifier CLIENT-IPV4-192.0.2.42
         Secret 6m9TXQTjLdH5BbgT
         RequireMessageAuthenticator
    </Client>
    

    Please note that configuring RequireMessageAuthenticator will require the Message-Authenticator to be present in the request and does not accept requests without it.

    A legacy RADIUS client, which is not able, or configured, to send Message-Authenticator is not able to connect with Radiator after this requirement. The RADIUS client should be configured to send Message-Authenticator in its requests and require Message-Authenticator in the replies. If only Radiator is configured to use Message-Authenticator but RADIUS client does not require it, an attacker can drop the Message-Authenticator from the reply to Access-Request and modify the contents of the reply.

    By default Radiator proxies and responds with Message-Authenticator to all messages with Message-Authenticator already present in them, but accepts also requests without the Message-Authenticator for compatibility and interoperability reasons.

    If your configuration uses SQL database to specify clients in addition to (or instead of) the Radiator configuration by utilising the ClientListSQL, please see ClientListSQL part of Radiator AAA reference manual. See documentation subsections GetClientQuery and ClientColumnDef for more detailed instructions.

    Requiring Message-Authenticator from RADIUS proxies or servers

    Radiator v4.29 (2024) adds RequireMessageAuthenticator configuration directive for RADIUS based AuthBys such as AuthBy RADIUS, ROUNDROBIN and HASHBALANCE. With RequireMessageAuthenticator enabled, the AuthBy only accepts requests with Message-Authenticator in the RADIUS replies. Radiator by default adds Message-Authenticator to outgoing messages.

    <AuthBy RADIUS>
         Identifier AUTHBY-RADIUS-PROXIES
         Secret 9iLeKAnBP8e8oMhb
         Asynchronous
         Retries 1
         RetryTimeout 3
         FailureBackoffTime 5
         # Requires Message-Authenticator from all RADIUS proxies/servers in this AuthBy
         RequireMessageAuthenticator
         <Host 192.0.2.111>
              # RequireMessageAuthenticator can also be required per host
              #RequireMessageAuthenticator
              AuthPort 1812
              AcctPort 1813
         </Host<
         <Host 192.0.2.112>
              # RequireMessageAuthenticator can also be required per host
              #RequireMessageAuthenticator
              AuthPort 1812
              AcctPort 1813
         </Host>       
    </AuthBy>
    

    Requiring Message-Authenticator in requests with Proxy-State attribute

    Radiator v4.29 introduces an additional configuration directive, LimitProxyState, which can be added to NAS RADIUS client definitions to require a valid Message-Authenticator in all requests, which contain the Proxy-State attribute used in the BlastRADIUS attack. This allows requests without Message-Authenticator (e.g. RADIUS client does not support it) to be received, but prevents the BlastRADIUS attack by not accepting requests with Proxy-State included in them. This configuration directive should only be used for NAS RADIUS clients, e.g network equipment such as switches, Wi-Fi controllers etc. It should not be used for clients, which are RADIUS servers or proxies. An example of the configuration is below:

    <Client 192.0.2.42>
         Identifier CLIENT-IPV4-192.0.2.42
         Secret 6m9TXQTjLdH5BbgT
         LimitProxyState
    </Client>
    

    Please note that RequireMessageAuthenticator should not be set together with LimitProxyState because RequireMessageAuthenticator will prevent requests without Message-Authenticator altogether. LimitProxyState is intended for use cases where client can not be configured to send Message-Authenticator.

    If your configuration uses SQL database to specify clients in addition to (or instead of) the Radiator configuration by utilising the ClientListSQL, please see ClientListSQL part of Radiator AAA reference manual. See documentation subsections GetClientQuery and ClientColumnDef for more detailed instructions.

    Limiting unencrypted RADIUS use only to management networks

    To utilise the vulnerability the attacker needs to get in between RADIUS servers or between RADIUS server and RADIUS client. Separating unencrypted RADIUS traffic into a management network and reducing the routing distance between the server and clients makes it more difficult to perform the actual attack.

    Unencrypted RADIUS over public networks exposes the RADIUS requests for the well-resourced attackers to abuse and should be avoided.

    Securing RADIUS messages with TLS or VPN

    The BlastRADIUS attack only works on unencrypted RADIUS traffic over UDP or TCP. The RADIUS traffic can already be secured with RADIUS over TLS (RadSec, RFC 6614) or with VPN solutions if RadSec support is not available.

    Radiator is the first RADIUS server ever (v3.12, 2005) to implement RADIUS over TLS neve(RadSec, RFC 6614), which was an IETF draft originally developed in cooperation between eduroam and Radiator implementers. Radiator Software continues to participate in developing the RADIUS over TLS specification in the IETF as well as implementing the latest features into the Radiator itself.

    Radiator Software strongly recommends migrating from unencrypted RADIUS to RADIUS over TLS. While requiring Message-Authenticator from RADIUS client and servers mitigates the vulnerability, the effort needed to configure these and test the changes, clients and servers for interoperability may be greater than configuring and deploying RADIUS over TLS to secure both the protocol and privacy of the information transferred in RADIUS request attributes.

    Questions and answers

    Is my multi-factor authentication (MFA/2FA) affected?

    If your multi-factor authentication is using RADIUS, it is likely to be affected. The exposure for the vulnerability depends on the RADIUS clients, servers and the attacker’s access to the network in between them. If your systems are for example doing MFA authentication with unencrypted RADIUS over the public Internet to a multi-factor authentication service provider your authentication is exposed and vulnerable. Is my enterprise Wi-Fi (WPA2/WPA3 Enterprise) affected?

    No. Enterprise Wi-Fi utilises EAP authentication over RADIUS. The EAP authentication already mandates Message-Authenticator in all requests.

    Is eduroam or OpenRoaming affected?

    No. Both eduroam and OpenRoaming are utilising EAP authentication like enterprise Wi-Fi (WPA2/WPA3) networks. OpenRoaming additionally mandates RadSec (RADIUS over TLS) and it is also supported in eduroam.

    Is my VPN solution affected?

    Similarly as with multi-factor authentication if your VPN endpoint is using unencrypted RADIUS for authenticating the VPN connection credentials from the RADIUS server or service, this vulnerability could be used in the worst case to allow an attacker to get a successful VPN connection to your network. Those VPN solutions which are using EAP authentication with RADIUS servers are not affected.

    Is roam.fi roaming service affected?

    The roam.fi roaming service is similar to eduroam and OpenRoaming utilising Enterprise Wi-Fi EAP authentication so it is not affected. Additionally roam.fi roaming service also supports RadSec (RADIUS over TLS) and it is strongly recommended that the roam.fi member organisations would migrate in using RadSec, when they have the RadSec connection capability themselves. The service will be updated to utilise Radiator v4.29 as soon as it becomes generally available.

    Is Radiator Auth.Fi RADIUS as a service affected?

    The Radiator Auth.Fi EAP authentication including both username-password and client certificate authentication is not affected.

    The captive portal and MAC address authentication used for guest network functionality is affected if unencrypted RADIUS connections are used. The Radiator Auth.Fi already supports RadSec client connections for securing the RADIUS traffic and it is strongly recommended that customers who have the ability to use RadSec, will switch to using RadSec instead of unencrypted RADIUS.

    The Radiator Auth.Fi Radiator will also be updated to utilise Radiator v4.29 as soon as it is released and that will enable additional options for adjusting Message-Authenticator settings for unencrypted RADIUS.

    Is my ISP, fixed line subscriber configuration provisioning or IP address allocation RADIUS solution affected?

    As these solutions are usually based on unencrypted RADIUS they are affected, but usually the RADIUS traffic is in these cases separated into management networks. If an attacker has access to a management network, the attacker is more likely to focus on more interesting targets than fabricating RADIUS requests and replies. The mitigation options work also in these networks, but extra attention should be given to testing, because the mitigation requires interoperable functionality from both RADIUS servers and clients.

    How can I update Radiator?

    If you have an active support contract for Radiator, you can get Radiator updates including the new Radiator v4.29 release from the Radiator download and repository page at https://radiatorsoftware.com/downloads/. If you are not sure or have an older license without a support contract, please contact sales@radiatorsoftware.com to renew your support.

    Is my RADIUS configuration affected?

    Radiator Software email support is able to answer questions and support you in case you want to evaluate the vulnerability’s effect on your deployment and use case. If you have active support, you can contact Radiator support at support@radiatorsoftware.com. If you are not sure or have an older license without a support contract or want to engage in larger scale configuration evaluation or upgrade, please contact sales@radiatorsoftware.com to renew your support or discuss the scope for Radiator expert services.

    For more information about the vulnerability

    BlastRADIUS WWW site: https://www.blastradius.fail/

    BlastRADIUS paper: https://www.blastradius.fail/pdf/radius.pdf

    CERT CVE: https://www.cve.org/CVERecord?id=CVE-2024-3596

    CERT Coordination Center: https://kb.cert.org/vuls/id/456537 (vendor status for vulnerability fixes)

    Radiator revision history (v4.29): https://radiatorsoftware.com/products/radiator/history/

    Wednesday, June 19, 2024

    How to speed up RADIUS authentication processing during peak hours with Radiator

    From time to time, our customers reach out to us and ask guidance on how to speed up RADIUS authentication processing during peak hours. Performance of this core functionality of Radiator is stable and predictable, and with increasing traffic load, most of the bottlenecks often emerge from the performance of the backend services that Radiator is utilising. In this blog, we give some helpful information to make improvements.

    What causes bottlenecks?

    In many cases,  authentication backends that Radiator is configured to use may be too slow to process the number of requests that are arriving. For example SQL databases may be slow to respond, LDAP lookups can take time, or HTTP requests be delayed. Writing accounting data or logging to a slow database can slow down Radiator.

    When this happens, the host system's UDP traffic buffer on the listening port fills up to the point that it overflows and new RADIUS request packets arriving at the server get dropped. Because the requests are dropped by the host operating system, they cannot even be logged as failed requests by Radiator.

    To get a better understanding of what is happening during peak hours, set `Trace 4` debug log level with `LogMicroseconds` in the configuration file (see [How to enable debug logging?]. The debug log will show how long each processing step is taking, and from there you can determine how many requests per second the configuration can handle.

    How to tune Radiator to speed up authentication processing?

    There are several ways to tune Radiator to speed up processing:

    • Split configuration into two separate files, one for authentication and one for accounting, and run Radiator as two separate running instances. This way there are two processes handling traffic with their own UDP buffers and backend dependencies that may even out the congestion. On Linux systems with systemd there are mechanisms readily available to help with that. More information: [How do I manage Radiator instances on a single host?].
    • Use a proxy instance to forward authentication and/or accounting requests to multiple worker instances (see first bullet). The requests can be spread out to worker instances so that request state is preserved (see [methods for load distribution and balancing] in documentation).
    • Within instances, use [FarmSize] to create a farm of server processes to add parallel processing of requests. Note: This is not compatible with many EAP protocols, such as EAP-TLS, EAP-TTLS, PEAP etc. This is because such protocols rely on authentication state being held within each server process, and it is necessary for all the requests for such protocols to go to the same Radiator process.
    • Incoming requests can be distributed from one or two proxying or load-balancing hosts to multiple hosts for processing.
    • Depending on use case, any combination of FarmSize, multiple instances per host and multiple hosts can be used together to enable more throughput. Separating processing on the Radiator side will create more separate connections to backends that may help even out backend load.
    Additionally, the bottleneck might be the available resources
    • If the host is running other processes, make sure it has enough resources available for Radiator. The best would be to use dedicated hosts for each function.
    • If a virtualisation platform is used, make sure that it can provide the configured resources to Radiator virtual machines even during the peak hours when other vm's are running their peak loads.
    • Logging to external servers may cause some unneeded delays for Radiator. This can be mitigated by configuring Radiator to log locally, and then process and ship logs with a separate agent.
    • Outside of Radiator, performance of backends may be tuned depending on the backend itself. For example a database can be scaled up or replicated to multiple-server cluster. On the other hand, speeding up the database might be possible even without adding more hardware if queries can be optimised or suitable indexes created. For optimising backends, please refer to their respective documentation.
    • Running configuration or Radiator may contain some parts that can be optimised. At the end of this blog post, we tell a bit about our consultation services. 
    How can we help?
     
    In case you would like our assistance when tuning your Radiator and related infrastructure, please contact sales@radiatorsoftware.com. Our team of experts have a long experience on configuring Radiator, and also assisting with the best practices when integrating Radiator with existing or new infrastructure.