Startup CloudFabrix has launched its Macaw chatbot, designed to bring generative AI to AIOps and upskill admin staff.
The ground is shifting beneath system maanagers’ feet as generative AI opens up a vast gulf between traditional manual-led system and application management and the new world of automation by accelerating AIOps (Artifical Intelligence Operations) of full system stack management. Truth be told there hasn’t been that much AI in AIOps – and what there is has been well-hidden. CloudFabrix wants to change that.
Shailesh Manjrekar, CloudFabrix VP of AI and SaaS marketing, said: “Generative AI is transforming the industry, but it is important to integrate it into the core product, understand the prompt context, and train on private data, which is uniquely done by CloudFabrix’s Macaw. This is going to fuel innovation for operational personas.”
Macaw is a generative AI assistant using natural language queries. It is claimed to uniquely identify the prompt context, leveraging CloudFabrix’s low code Robotic Data Automation Fabric (RDAF) platform. It then uses both Azure’s OpenAI LLM and CloudFabrix’s LLM for semantic search on the customer’s local body of knowledge to glean insights and investigate data, compose and explain pipelines, and compose dashboards and service tickets.
In order to prevent leaking information and ensure it operates on the customer’s dataset, CloudFabrix says Macaw ensures privacy and governance over local data and data transfer between large language models (LLMs) with user-defined policies.
This CloudFabrix release also includes RDAF Edge, which is packaged in a single VM to run on edge endpoints. It’s aimed at 5G and edge use cases for telcos and other customers, and ingests data with real-time topology enrichment, sending it to a central data platform or data lake.
CloudFabrix says telcos need to ingest a large number of different data types such as syslogs, SNMP traps, gNMI, Bulkstats, and OpenTelemetry across domains – Campus, Datacenter, Cloud, Mobility/5G, Optical, etc. This data can come from devices, controllers, physical, virtual and cloud-native network functions. Telco Service Assurance functions need a single pane of glass to visualize aggregated data across these disparate sources with persona-based composable dashboards.
Rather than coding a fixed set of reports abased on API hooks into the data-generating infrastructure, CloudFabrix has designed its Macaw chatbot to “talk” to admins and then compose dataset queries on top of microservices and event-driven Functions-as-a-Service. There is also pre-built code for composability, field customization, and extensibility – all of which CloudFabrix says lends itself very well to multiple data-centric use cases.
CloudFabrix says Macaw breaks down operational data silos and should spread observability and AIOps more widely. As a first phase of this, Macaw will be used for upskilling and reskilling of operational roles and subsequently for AIOps insights.
Many incumbent storage providers have predictive system management monitoring and analytics delivered from the cloud and based on telemetry from customers arrays. HPE’s InfoSight is a classic example. Pure’s Pure1 operational management facility is another.
The product managers and engineers responsible for these will all be looking at how chatbot technology could be used to improve their operators’ ability to monitor, manage and fix problems in their products.