Trellix costs lowing costs, enlightening speed, and adds service delivery with Amazon Nova Lite Lite Lite Lite Lite Lite Lite Lite Lite Lite Lite Lite Lite Lite Lite Lite Lite Lite Lite Lite Lite Lite Lite Lite Lite Lite Lite Lite Lite

This post is written by Martin Holste from Trellix.

Social groups are dealing with the exposure of cyberercere threatening. These threats are increasing in the form form, and the attack area focuses on. To be forced by the estimate of talent and budget, groups that are often forced to prioritize events under investigation, reduce the power to find new threats. Trellix Wise powerful AI enables security groups to convert the threat to threaten and cannot risk scores in events. With smellix intelligent, we can now complete the use of many hours of work in secondaries, allow them to increase your ability to.

Trellix, the best company releases a broad-wide range of a broad extensive Ai-Powered Ai-Power. The full-time, open, open, and traditional AI programs assist organizations make up working against advanced threats. Trellix Wise is available to customers as part of the Trellix Safety Platform. This post discusses the acquisition and assessment of Amazon Nova Foundation Models (FMS) by TRellix.

With growing acceptance and use, the trellix group has been exploring ways to expand the cost of a logical investigation. Small FMS, which is very expensive that they promise and amazon nova micro emissions as a result of their quality and cost. In the early tests, the trellix group realized that Amazon Nova micro delivered three times and under 100 costs.

The following statistics are the results of tests by Trellix by comparing amazon Nova micro models in Amazon Bedrock.

TRELLIX group has identified areas where Amazon Nova Micro can complete their Anthropic's Claudes Sonnet use, bring low cost and high speed. In addition, the TRELLIX qualified team received Amazon Nova Lite to become a strong model and understanding code and now using Amazon Nova Lite to accelerate their custom service.

Trellix Wise, Based on a Powerful threat to help safety commentators

TRellix is ​​wise in Amazon Bedrock and using anthropic's Claude Sonnet as its main model. The platform uses Amazon Opensech service stores billions of security events collected from the underlying areas. The OPENSearch Service comes with the power of the formation of the built-in vector, making it directly to use the data stored on the Openson Generior of Amazon Bedrock generation. Using Opensearch and Amazon Bedrock service, Trellix Wisdom serves measures to investigate the default, braceleting at each party. This includes the return of the data to analyze, analyze data using Understanding from other customary (ML) learning models (ML), and risk factors. This complex approach enables the app to translate complex security data patterns and make wise decisions for each event. The smellix investigation of the trellix provides each event for each risk and allows analysts to enter the most deeper in the results of analysis, determining that following people are required.

The following screen displays an instance of an event in Trellix Wise Dashboard.

With the growth of the approval measure, Trellix has been investigating costs improve costs and speed. TRellix team decided that all investigations require the accuracy of Claude Sonnet, and that some stages can benefit from the expensive, low-income models. This is where Amazon Nova micro has helped improve the cost of inquiry.

Improving an investigation costs with Amazon Nova Micro, Rag, and repeating

The threatening investigation is consisting of many measures, from data collection, processing, assigning an accident scaret for the event. The collection stage returns the event related information to analyze. This is done using one or more drop-down calls in the model in Amazon Bedrock. The main in this section increases the completeness of return data and reduces incorrect (hallucinations). TRELLX team pointed to this section as a relevant phase in the work of work to use speed and cost.

Trellix's group is finished, based on their testing, Amazon Nova Micro offer two important benefits. Its speed allows you to process 3-5 to look at the same time as one-day Sonnet infant and costly expenses approximately 100 times. TRellix team decided that using many calls, you can increase the coverage of the required information and is still low costs in a period of 30. Although the model answers have higher variables, a lot of passing empowering the complete response. Response Restrictions Unemployment Terms Terminal Development Development and Reference Data Develop response, Halkinacinaciations and Introduction to Response.

Before using this method, the Trellix team has made detailed testing for perfection, expense, and speed. The team recognizes the beginning of their AI-producing trip where the general benches are not enough when assessing the models for a specific case of use. The multiplication testing evaluation of jobs was established and a detailed examination of many models, to ensure the benefits of this method before moving forward. Speed ​​and cost Benefits seen by Trellix have helped to ensure benefits before submitting a new productive method. This method is now sent to the limited driving area. The detailed test is done as part of the phase creation.

Store

In this post, we share how TRellix welcomed and explored the Amazon Nova models, which resulted in the greater thief softens and low cost. The project is, a Trellix team recognizes the following as level donors will allow them to achieve these results:

• Access to various models, including small skilled models such as Amazon Nova Lite and Amazon Nova Lite, accelerating the ability of a group of easy assessment and accept new models.
• The power to force answers to avoid beauty, using pre-operated system entries for information concerning information, procedures, and policies, reduced the risk of understanding and inadequate.
• Data services that have enabled effective integration of data by the simplified model models and reduced the time to produce new components.

“Amazon Bedrock makes it easy to explore new models and methods as available. Using Amazon Nova Sonnet's Othropic's Claude.” Said Martin Holste, the Executive Director, Engineering , Trellix. “We are truly happy with flexibility by Amazon Bedrock allows us to continue to check and improve smellix trellix and trellix security platform.”

Get started with Amazon Nova at Amazon Bedrock Console. Learn more on the Amazon Nova product product page.

About the authors

Martin Holste CTO of cloud and Gyanya Etrellix.
Firat Elbey It is the main product manager in Amazon Agi.
Deepak Mohan by the main product manager of AWS.