Scylla Summit 2019

Scylla enables rapid Machine Learning experimentation at Tubi. The current-generation personalization service, Ranking Service, ramps up experimentation by 5x, while Popper, the next-generation experimentation engine, will grow by 10x and beyond. We'll talk about what's so special about these services.

Scylla includes multiple features that collectively provide a robust security model. Most recently we announced support for encryption-at-rest in Scylla Enterprise. This enables you to lock-down your data even in multi-tenant and hybrid deployments of Scylla. Join Tzach and Dejan for an overview of security in Scylla and to see how you can approach it holistically using the array of Scylla capabilities. He will review Scylla Security features, from basic to more advanced, including:

• Reducing your attack surface 
• Authorization & Authentication 
• Role-Based Access Control 
• Encryption at Transit 
• Encryption at Rest, in 2019.1.1 and beyond 

LDAP authentication is a common requirement for any enterprise software. It gives users consistent login procedures across multiple components of the IT infrastructure, while centralizing the control of access rights. Scylla Enterprise now supports authentication via LDAP. We will look into how to configure Scylla Enterprise for LDAP interaction and how to fine-tune access control through it.

Augury stores and serves time-series features from massive streams of IoT data, both for real-time insights, and offline learning and analytics. Learn about Augury’s needs and constraints, their solution evaluation and architecture, and fundamental practices for efficient data modeling, plus get a glimpse into the next-gen architecture at Augury, with a view on time-series feature storage and serving.

The Scylla Cloud is a flexible service, offering many options for its users on how to configure their deployments. Clusters can be spawn into one or multiple datacenters, with VPC peering and in either Scylla's or the user's account. This talk will explore all the options available for users of the Scylla Cloud and how to make the best of each of them, as well as explore what's coming in the next 12 months.

FireEye believes in intelligence-driven cyber security. Their legacy system used PostgreSQL with a custom graph database system to store and facilitate analysis of threat intelligence data. As their user base increased they ran into scaling issues requiring a system redesign with a new platform.

This presentation will focus on the backend systems and migration path to a new technology stack using JanusGraph running on top of Scylla plus Elasticsearch.

Using Scylladb turned out to be a game-changer in terms of performance and the types of analysis our application is able to do effortlessly.

iFood is the largest Brazilian-based food delivery app company. It connects users, restaurants, and deliverymen using an event-driven architecture using AWS SQS and SNS, with programming in Java and Node.js. Thales' team is responsible for delivering orders' events to restaurant devices at least once, which is currently done using a REST API polling and acknowledgment system.
Learn how their database infrastructure evolved from a PostgreSQL database, but began to show limitations and was a single point of failure. Growing through a few intermediary steps, including Amazon DynamoDB, eventually, turning to Scylla for its data model and collections to condense multiple tables. Using Scylla, iFood reduced the time to process events and acknowledgments (from ~80ms to ~3ms) and reduced costs using Scylla vs DynamoDB by over 9x.

Nauto provides an intelligent driver safety solution and fleet management safety application that increases fleet safety. The talk will explain two infrastructural components: online driver identification and trip generation (with the route). This talk examines the algorithms for each and how Nauto uses Scylla as the backing data store to do this at scale.

Most interestingly, you will learn how Nauto guarantees consistency by modeling a section of their algorithm as a function that is atomic, commutative, and idempotent (ACID 2.0).

Project Gemini - Roy Dahan, ScyllaDB
Bugs are a fact of life, and often they only happen under certain circumstances. It could be that an issue only appears with a certain number of columns, only when blobs reach a certain size, or only when TTLs expire too far into the future or too close to the current date. With a number of possibilities that is exponential in nature, how to catch them all? This talk will explore Project Gemini - a fuzzing tool used by Scylla to guarantee that data, once written, is always safe and sound.

How We Make Adding and Removing Nodes Faster and Safer - Asias He, ScyllaDB
When a new node is added or removed, Scylla has to transfer part of the existing data from some nodes to their neighbors. When a node fails, Scylla has to repopulate its data with data from the surviving replicas. Those operations are collectively referred to as "streaming" operations, since they simply stream data from one node to another, without using this opportunity to also fix discrepancies in the data. This is in contrast with the repair operation, that looks into all existing replicas and reconcile their contents. Scylla is moving towards unifying those two operations. In this talk we will discuss why this is considered beneficial, and what other possibilities this opens to users.

Scylla now supports a choice between Global and Local Secondary Indexes. Their performance characteristics will be widely different, and better or worse depending on the situation. Which will perform better, when and why? We will also discuss filtering of data, a technique to allow the server to select just parts of the data before sending it to client, yielding faster queries. What can be filtered, and in which situations will it yield a performance gain? How can we use indexes to speed up the performance of filtered queries? Those and other questions will be explored in this talk. 

Event-driven architectures are increasingly part of a complete data transformation solution. Learn how to employ Apache Kafka, Cloud Native Computing Foundation’s NATS, Amazon SQS, or other message queueing technologies. This talks covers the details of each, their advantages and disadvantages and how to select the best for your company’s needs.

An online sports apparel powerhouse, Fanatics has to power an always-on scalable ecommerce site listing wares for over 1,000 vendors, and to track the shopping carts of millions of demanding sports fans. Find out how Fanatics uses Scylla for order capture, how they handle abandonments, perform data mining and more. Learn their “practical engineering” lessons, and discover how Fanatics wrote their own management tool for cluster automation and autoscaling on AWS.

How can Kubernetes be best used to automate the deployment, scaling, and various operations of a Scylla database?
Enter Kubernetes Operators, the way to combine domain-specific knowledge about Scylla with the automation framework of Kubernetes.

In this presentation, we will quickly explore what Kubernetes is and why it works so well, highlight the pain points of running Scylla with just Kubernetes primitives, and show how we extended Kubernetes so that it can correctly operate a Scylla database.

Finally, we will show the Scylla Operator in action and show how easily you can spin up a Scylla cluster with just one command.

Are you running separate database clusters for operational and analytical workloads? Scylla now has the ability to handle multiple workloads from a single cluster--without performance degradation to either. This session will cover:

• The evolving requirements for operational (OLTP) and analytics (OLAP) workloads in the modern datacenter 
• How Scylla provides built-in control over workload priority and makes it easy for administrators to configure workload priorities 
• The TCO impact of minimizing integrations and maintenance tasks, while also shrinking the datacenter footprint and maximizing utilization Plus we’ll share test results of how it performs in real-world settings.