Being data-driven is extremely crucial for any organization looking to enable continuous innovation, to avoid falling behind and to lower TCO. Much like DevOps in the enterprise, the emergence of enterprise DataOps mimics the practices of modern data management at large internet companies over the past 10 years.
DataOps is about automatically managing all your data life cycle, not necessarily related to machine learning. It’s about how to provision the resource for your data, scale the database, tune the data for performance, and automate the data management or data team between different platforms or different stores.
The GigaOps Stack reduces the complexity of all aspects required to implement GigaSpaces products via a self-service cluster with end-to-end provisioning, ability to deploy new services, auto-recovery, self-monitoring, and tracing.
The GigaOps Stack has the following advantages when compared to Kubernetes:
- Better performance for GigaSpaces products - up to 20% faster
- Broader installation options - supports VMs in addition to Docker containers
- Supports more environments - can be used in hybrid environments that include both cloud and on premise
- Simplified networking and security configuration
- Lower cost for worker nodes (GKE/EKS/AKS)
The GigaOps Stack is comprised of the following:
- gsctl - a standalone CLI tool that you can use to create, provision, and install GigaSpaces clusters in the cloud or on premise.
- Ops Manager - the GigaSpaces native administration and monitoring tool, which is web based and runs automatically when GigaSpaces services are deployed.
- Monitoring stack - based on InfluxDB, Grafana, and Telegraf for monitoring GigaSpaces cluster activity and metrics using dashboards
- ElasticGrid or Kubernetes orchestration
You can monitor and administer your environment using Ops Manager, and ElasticGrid includes pre-defined Grafana dashboards for self-monitoring.
The GigaOps Stack utilizes the DataOps approach, which takes DevOps best practices and applies them to managing data as a global asset in your organization. Using the GigaOps Stack provides auto-installation and provisioning of servers, networking, and security using gsctl, along with full orchestration and installing the GigaSpaces platform.
With the GigaOps Stack, the installation and configuration process is completely automated. This saves time and resources, and prevents errors that might otherwise occur due to manual configuration of the environment.
The GigaOps Stack manages orchestration, scheduling, service life cycle, and resource prioritization across the environment. The stack makes global decisions about scheduling, and can detect and respond to cluster events. For example, identifying healthy and unhealthy nodes so that it can make decisions about where to place pods.
The GigaOps Stack supports canary deployment, so you can apply GigaSpaces upgrades or custom application upgrades to a small subset of users before introducing the change on a widespread scale.
As the load on your system gets larger or smaller, you may want the size of your cluster to adjust accordingly, to reduce spending on cloud services when they aren't necessary, or to maintain performance during peak times. With GigaOps Stack, your cluster can scale up or down automatically to support the current load.
You can implement GigaSpaces product upgrades or your own applications across your cluster with zero downtime.
Benefit from high availability and robustness with the GigaOps Stack's ability to self-repair if a cluster node becomes unavailable or starts to fail.
ElasticGrid comes with pre-defined Grafana dashboards for monitoring your GigaSpaces cluster. You can use these dashboards in conjunction with canary deployment to monitor performance and other metrics for each version that is deployed, or to assess the impact of a configuration change by comparing the metrics before and after it is implemented.
ElasticGrid includes Zipkin, a distributed tracing system that gathers timing data to enable troubleshooting latency issues in the system. This enables more effective performance tuning.