Cloud integrated storage (CIS) is a relatively new offering for Open Enterprise Server. CIS can provide an organization with many benefits including:
- Data insights from all the OES servers to identify hot and cold data over a configurable time period
- Secure gateway to push, manage, and access data across a private or public cloud
CIS can be configured in multiple ways, but to reap the benefits fully, it is important to understand which configuration option is right for your current circumstances. With the correct configuration option, organizations will see faster migrations and quicker access to migrated files.
CIS Architecture
Before diving into the various configuration types, it is helpful to understand the components that make up a CIS server. These components can co-exist on the same physical server or can exist across multiple servers and make up a single CIS deployment. The multiple configurations are how the components of CIS can be installed and configured across multiple servers to provide high availability, load balancing, and increased reliability.
Database
CIS works with either an external database (MS SQL or Maria DB) or an internal MariaDB database that comes bundles with OES servers. The database is used to store information about the OES servers, cloud accounts, policies, and migrated metadata.
CIS Infrastructure Services
There are various third-party applications leveraged by CIS. For example, it utilizes Apache Zookeeper as the configuration store and is used by all CIS configurations deployed in your organization. Apache Kafka is used as the messaging system across services. Elasticsearch is used for analyzing the hot and cold data in your organization. It is also used in Dashboards where you can see how data been moved to the cloud or recalled.
The infrastructure services can be installed as an independent cluster to ensure high availability but are also part of a single CIS server configuration. For a single CIS server, the infrastructure services do not run in any cluster mode, reducing reliability.
CIS Services
CIS core services form the fundamental building blocks of a CIS server. Included are the main services, authorization, policy, insights, dashboard enablers, and the management server. There can only be a single instance of CIS core services running in your organization and to ensure services are not disrupted, it is recommended to run CIS services as an OES cluster resource.
CIS Data Service and CIS Gateway Service
CIS Data service is the CIS component that connects to the cloud provider. It can migrate data to the cloud from an OES server and recall files from the cloud to an OES server. To accomplish this, a CIS data service is configured to work with a CIS Gateway service.
A CIS Gateway service provides an entry point to all the CIS services on that server and can be an entry point to many CIS Data services, acting as a load balancer between them.
There can be multiple CIS Data services running in an organization to improve the latency of file access from the cloud. An independent CIS Data service can be the Data service for multiple OES servers and are remote from the main CIS server. Multiple Data services can connect to a single Gateway server and multiple Gateway services in an organization can connect to the main CIS server.
On a single CIS server, there is an instance of Data service and Gateway service running. To ensure CIS services remain not disrupted, it is recommended to run the Gateway service as a cluster resource.
How to Know Which CIS Deployment is Right for You
There are multiple configuration options for your CIS deployment:
- Basic
- Highly available Infrastructure services + CIS Core Services (single or cluster resource)
- Highly available Infrastructure services + CIS Core Services (single or cluster resource) + CIS Gateway service (single or cluster resource) + CIS Data service
Let’s look at which one would be right for you.
Proof of Concept (Pilot)
This option is for organizations looking to evaluate CIS.
Since the objective is to first evaluate CIS for your organization, you’re going to want a simple configuration so CIS is ready to be used as a pilot in just a few clicks. The single server deployment configuration will be suitable as it configures the Infrastructure service, core CIS service, Data service, and the Gateway service on one server.
This setup is only recommended for this pilot scenario. The database can be local or remote. You will likely want a local database if the pilot data is small enough and if you do not want to host it on the database server in your organization. If you can host the database on an existing database server in your organization, you can choose the external database option. CIS supports both MariaDB and MSSQL databases.
Local OES Servers With Fewer Users
This option is for organizations that are medium-sized with all OES servers in a single location and not many branch offices.
For this case, it is recommended to have the following CIS deployment:
- External database with replicas
- Highly available Infrastructure services
- CIS services as an OES cluster resource
All the OES servers in your organization will connect to the CIS server’s Gateway service and all data migration to and recalls from the cloud will be managed by the local CIS Gateway service.
Local OES Servers With Many Users
This option is for organizations medium sized with all OES servers in a single location and many users who connect to a few of your OES servers for their data. These OES servers see more traffic than the other servers in your organization.
For this case, it is recommended to have the following CIS deployment:
- External database with replicas
- Highly available Infrastructure services
- CIS services as an OES cluster resource
- CIS Data service on the busy OES servers.
- CIS Gateway services on one of the OES servers, where Data service is configured.
All the OES servers where the Data service has been configured will connect to the independent CIS Gateway server deployments. These Gateway servers will be configured to connect to the single CIS Core server in your organization. Having multiple Data services connected to a single Gateway service will ensure load balancing between the two. Additionally, placing the Data services closer to where the data access is will improve latency.
Remote OES Servers With Many Users
This option is for large organizations with OES deployments spread across several locations. Some of these OES servers might cater to more users than the other OES servers in that location.
For this case, it is recommended to have the following CIS deployment:
- External database with replicas
- Highly available Infrastructure services
- CIS services as an OES cluster resource
- Multiple CIS Data Gateway services spread across the locations, with each gateway server proxying multiple data servers within that location.
If you have large branch offices spread across the same remote location, you can configure data servers in each of the branch offices, with a data Gateway server in the remote location. This allows for multiple data gateway servers that all connect to the core CIS server.
Choosing the right CIS deployment is key to providing near-zero latency for user file access in the cloud. Migrations are also faster, which means your data reaches the cloud faster with built-in load-balancing features, providing your organization with reliability and high performance.