Redis Revolutionizes GenAI with Seamless Real-Time Data Empowerment


CXOToday has engaged in an exclusive interview with Yiftach Shoolman, Co-Founder & CTO, Redis


Ques: What are the notable features introduced in Redis version 7.2, particularly in the context of data integration and real-time serving?

Ans: Redis version 7.2 introduced an innovative feature: Redis data integration approximately six months ago. This functionality facilitates the smooth transfer of data from diverse databases into Redis, enhancing real-time serving capabilities. Redis has a distinguished history in the field of AI, pioneering the integration of inferencing in 2018, subsequently adding feature store capabilities, and establishing a vector database for recommendation engines. Presently, our suite of Gen AI capabilities empowers applications to access pertinent, up-to-date, and secure data.

There has been a notable surge in both cloud and on-premise usage, underscoring Redis’s growing popularity. A significant focus has been placed on the development of globally distributed databases strategically positioned for proximity to end-users. This strategic placement ensures a real-time experience, accompanied by a commitment to achieving uptime levels of five nines and above. This commitment involves compensating customers for any outage exceeding 26 seconds per month, a crucial aspect given the substantial volume of operations per second.

Anticipating industry consolidation in the database space, particularly with approximately 350 companies vying against the top 10, Redis is at the forefront of this transformative landscape. Our technology places a strong emphasis on cost-effectiveness for real-time use cases, resulting in a notable reduction in customer expenses by 40 to 60% in cloud deployments. Managing millions of operations per second in real-time databases is inherently challenging, often leading to overprovisioning and unforeseen costs. This overview underscores Redis’s unwavering commitment to innovation and its proactive approach to addressing prevailing trends in the dynamic business environment.

Ques: How does Redis streamline the integration process for model developers with platforms such as OpenAI, AWS Bedrock, Google Vertex API, LMM Chain, and Lambda Indexes, ensuring a smooth and user-friendly experience for developers?

Ans: For model developers, post-integration focuses on platforms like OpenAI (via direct offerings, Azure OpenAI API, AWS Bedrock, Google Vertex API, or open-source tools like LangChain or Llamaindex), all seamlessly interfacing with Redis. We recently announced that OpenGPT service by LangChain integrates successfully with Redis. The goal is a frictionless experience, with Redis handling everything associated with the real-time data of GenAI while developers utilize platforms seamlessly.

Redis provides a full suite of solutions for accessing GenAI data in real-time:

  1. Our vector search capabilities provide fast access to the knowledge base. Based on our benchmarks, Redis is by far the fastest vector search on the market
  2. By providing access to Conversation History (some call it LLM memory), you can now enrich the LLM with relevant data and improve accuracy. It improves the quality of your GenAI app and reduces hallucinations

3 A Semantic Cache enables you to respond directly to questions from promt that are semantically identical (e.g., “how many people live in India,” “what is the population of India”), significantly improving the response time of your GenAI app and reducing infrastructure costs through the elimination of LLM calls.

Ques: What distinguishes Redis Cloud solution from AWS, especially in the domains of AI and database technologies? How does Redis’ serverless, coupled with superior performance benchmarks and cost savings, position it as a compelling choice for users seeking advanced, efficient, and cost-effective solutions compared to AWS?

Ans: Artificial intelligence was prominently featured at AWS Reinvent, dominating booths and captivating audiences. Despite some skepticism, major tech firms are heavily investing, with plans for deploying 100 GenAI applications next year. The undeniable impact of AI highlights the crucial significance of intellectual property (IP).

Redis Cloud stands out with its serverless solution, simplifying user experiences without the complexity of managing clusters. Notably, the key differentiators lie in (1) uptime, Redis Cloud provides 5 nines versus four nines uptime of ElastiCache; (2) Redis Cloud can run globally across regions, clouds and hybrid deployments and serve the data closer to where the end users are, with A-A technology, which AWS ElatiCcahe does not support; (3) Redis Cloud supports Search and Vector Search capabilities, making the GenAI application faster, and allowing users to search in their Redis datasets, ElastiCache does not support this; and  (4) TCO—Redis Cloud offers a replicated gigabyte at $20-$40/GB/mo compared to $90 of AWS ElatiCache. ,

When comparing performance and scalability:Redis Cloud’s benchmarking in 2019 showcased superior performance, achieving 5 million operations per second per instance compared to AWS’s 1 million (on much newer instances.. The architecture, while not delving into specifics, positions Redis Cloud to outperform AWS by leveraging technology that is genuinely 10 times better.

Competing with AWS necessitates superior technology. Redis believes its technology meets this criterion, which is evident in its ability to offer advanced capabilities at a better cost without compromising performance.

Ques: Could you provide details about some of the customers in India?

Ans: Let me share information on three Redis customers and how Redis has helped them. Apna, a platform catering to job seekers and employers in tier two and tier three markets in India, functions as a LinkedIn for blue-collar workers. Through Redis, Apna enhances job seekers’ profiles, providing real-time, personalized job recommendations based on location, experience, and availability—an impactful use case. Another aspect involves utilizing AI to assist first-time internet users in articulating their skills through automated sentences, thereby improving their employability.

Moving on, Zee Entertainment initially employed AWS but encountered challenges managing data across five regions. Upon transitioning to Redis Enterprise, they embraced the active-active feature, facilitating automatic updates across regions and significantly enhancing operational efficiency. This shift resulted in an impressive 70% cost saving, serving as a success story that underscores the advantages of Redis Enterprise.

Lastly, a trading platform underwent a transformative shift with Redis Enterprise. Before adopting Redis, they grappled with latency issues at 1 million operations per second, resulting in 480 milliseconds. Following migration, the latency dramatically decreased to just 20 milliseconds—an outstanding improvement from 1/2 to 1/50 of a second. These stories vividly showcase the diverse and impactful applications of Redis technology.

Ques: What does the future look like for Redis?

Ans: We anticipate a positive and promising future for Redis. When examining trends that influence the definition of real-time, such as network latency, device performance, the shift from two-dimensional to three-dimensional interfaces, immersive experiences, and the integration of AI, it becomes apparent that the factors driving the need for a real-time database are converging simultaneously and evolving rapidly. This convergence of trends is highly favorable for organizations closely aligned with Redis, positioning them to thrive in an environment characterized by dynamic change.