In the realm of speeding up data access, choosing the right in-memory data repository is crucial for optimal application performance, change management and platform flexibility. Among the various options available, Redis, as a representation of advanced caching solution, and GigaSpaces XAP Skyline, a top notch In-Memory-Data-Grid (IMDG), are viable alternatives. Those two stand out as prominent contenders in their domain. This differentiation analysis aims to shed light on the distinct characteristics, use cases, and capabilities of these two technologies.

Redis: A Versatile In-Memory Cache / Unstructured Data Store

Redis is renowned for its versatility. It can be employed as a cache, a NoSQL database, or a message broker, catering to a wide range of application needs. Its high performance and scalability make it a popular choice for application caching scenarios.
Redis offers flexible schema design, horizontal scalability, and high availability. It is used to store unstructured data such as documents and images flexibly. Unlike relational databases, it allows storing data in dynamic models that are schemaless. Which means that data needs to be prepared and modeled per expected query prior to storing it, in case different representations are needed, it will need to be stored several times in different formats.

XAP Skyline: A market leader IMDG, Engineered for Large-Scale Data Management

GigaSpaces XAP Skyline provides extreme processing and represents a specialized category of storing both documents and structured objects-based data in-memory. It is, in fact, a distributed database specifically designed to handle vast amounts of data in memory. Its primary objective is to accelerate data access while assuring consistency thanks to its ACID compliance and SQL support that makes it easy to interact with.

Upon write operations, XAP Skyline supports write behind scenarios so that it can synchronize a backend RDS with data written to it, thus keeping cache and backend DB in constant sync.

XAP Skyline, unlike caching solutions such as Redis, also provides high performance processing capabilities that run parallel distributed computations under extreme loads. Its built-in events mechanism, and the ability to execute business logic using “processing units” that are co-located where the data nodes hold their data, provide additional OOTB capabilities to support additional extreme scale scenarios. 

Use Cases and Suitability

The suitability of Redis as a data store is evident in applications that require:

• Low latency and rapid data access
• High scalability to handle increasing concurrent requests
• Support for various complex data types, including strings, hashes, lists, and sets
• Support for dynamic semi-structured datasets

On the other hand, XAP Skyline is ideally suited for applications that demand:

• Predictable ultra-low latency, as the data is not lazy loaded but prefetched.
• High data throughput also for processing large volumes of data efficiently
• In-memory data repository  of substantial structured and unstructured datasets for faster access
• Compatibility with SQL queries for easier data retrieval
• Write-behind capabilities to enhance write performance and assure consistency.
• Automated prefetching of data from persistent storage to memory for improved read performance
• Support for structured indexed data, support for flexible data retrieval without duplicating the data per each query.

Data Access and Storage

Redis offers a RESTful API for accessing data stored within its in-memory technology data structures. This API provides a simple and efficient way to interact with Redis, enabling developers to leverage its capabilities seamlessly in their applications. Developers also need to manage the Redis cached data lifecycle, it’s the application responsibility to store the data & check whether it’s available in the cache. Data retention can be done automatically per TTL (time-to-live) parameters.

XAP Skyline in-memory technology is ideal as a data infrastructure level performance tier, capable of serving multiple applications that require high data throughput, low-latency, unstructured data management, with predictable performance and consistency, and SQL-like query capabilities. It is also a great fit in cases where part of the application logic can benefit from residing inside the performance tier, potentially even in the same memory space where the hot data resides.  

One of the key differentiators is that XAP Skyline employs an initial load of the data, and then as data update operations are performed through XAP Skyline, data freshness is guaranteed and further real-time data processing can take place as needed. Data is stored in indexed tables/objects, while with Redis it’s more about lazy loading of data, which is then stored as an unstructured entity. This data is not queryable through flexible mechanisms such as standard SQL, and becomes invalidated when its time-to-live (TTL) window expires.

XAP Skyline provides both RESTful APIs and SQL access to the data it manages. As it implements the PostgreSQL protocol, any application with a PostgreSQL driver can access XAP Skyline as if it was an RDS.

The SQL support allows developers to use familiar SQL queries to retrieve, manipulate, and analyze data stored in XAP Skyline. As an IMDG, it simplifies data access and reduces the learning curve. It also allows the introduction of a seamless performance tier between the application and data tier. Unlike a transient cache, it minimizes changes to the application changes while it can also guarantee data freshness.

XAP Skyline stores the data both in RAM and on a tiered storage, local SSD. This mechanism helps organizations better control their cloud spending, as RAM can become quite expensive when storing large amounts of data.The platform provides a built-in event-driven-architecture (EDA) system that triggers custom business logic per data change events. This enables high performance, on the fly data processing inside XAP Skyline, where business logic requires extremely fast data access.

Accelerating Data Access in Hybrid & Cloud Environments

Both Redis and XAP Skyline can be deployed in cloud environments to accelerate data access for applications running also on Kubernetes. 

Redis has become a default choice for cloud-native applications that require simple data access with minimized latency. An application that uses Redis is responsible for managing the data in the cache. When it queries for data, it needs to check whether the data exists in cache, if so – fetch it from cache. If not – get it from the backend system and store it in the cache for future use, in the format fit for that particular use. 

Data will not be refreshed; it will be invalidated per its TTL policy. This pattern creates both unpredictable response time, as the cache hit ratio does not stand at 100%, and the consistency is also not guaranteed, as data in the cache is not fresh and does not get updated.

XAP Skyline changes this approach and is situated between the managed RDBMS and the  cloud native application, preferably even in the same Kubernetes cluster. This avoids the network  latency introduced by cloud network hops that are naturally added by Kubernetes and from the network topology of managed data services. 

Cloud-native applications can continue using SQL queries just like they used those queries before the introduction of XAP Skyline. The data is served from XAP Skyline, as it was already pre-fetched. When the application updates the data on XAP Skyline, the platform performs the write back to the managed RDBMS while it assures consistency, as it is ACID compliant (Transactional). It writes the updated data back to the managed RDBMS, thus replacing the RDBMS for data that needs high concurrency, low latency access from the application’s perspective.

Conclusion

Redis and XAP Skyline are powerful data repositories that offer unique capabilities for enhancing application performance. The choice between these two technologies depends on the specific requirements and characteristics of the application. For use-cases that prioritize application level caching via simple deployment that is tightly coupled to the application, Redis is a good option. 

Is there a Redis alternative? Yes, XAP Skyline is ideal as a data infrastructure level performance tier, serving several applications that require high data throughput, low-latency, true data management, predictable performance and consistency, and SQL-like query capabilities.