Highlights
● A powerful, scalable, elastic in-memory grid for your business-critical applications
● Offers broad range of usage scenarios spanning from in-process cache to powerful distributed grid
● Lowers loss risk via automatic data replication, delivering high availability and fault tolerance
● Ability to quickly and easily increase cache capacity and data throughput as needs grow
● Flexible and simple user management interface for administration and monitoring
IBM WebSphere
Distributed Caching
Products
Today’s dynamic business environment and economic uncertainty mean organizations must work smarter to remain competitive and respond to changing customer demands. The key to working smarter is business agility and cost optimization. Our planet is becoming increasingly interconnected, instrumented and intelligent. As more people around the planet join the digital world, they drive exponential growth of busi-ness and non-busibusi-ness transactions for everything from web commerce and RFID processing in dynamic supply chains to real-time financial interactions to online social networking and gaming. Organizations need to ensure that their critical applications can meet the require-ments of rapid increases in demand, can deliver immediate, consistent responses and can scale as necessary. Additionally, there is a need to enhance application performance to meet rising customer expectations around application response times while managing the costs of your infrastructure.
The IBM WebSphere® portfolio offers extreme transaction processing (XTP) capabilities that ensure you have a SMART SOA™-based application infrastructure, which can support your most demanding business-critical applications. WebSphere improves your business agility by allowing you to overcome traditional IT performance limita-tions to generate the levels of global scale, process efficiencies and business intelligence needed for smarter business outcomes, like sus-tainable competitive advantage, maximized revenues and avoidance of potential fines stemming from inconsistent response and Service Lifecycle Agreement violations.
Additionally, the IBM WebSphere caching family of products allows business applications to process billions of transactions per day with extreme efficiency and near-linear scalability. They are designed to work in heterogeneous environments across all leading application server platforms and virtualiza-tion environments.
The IBM WebSphere eXtreme Scale and
WebSphere DataPower® XC10 Appliance
provide distributed object caching that
is essential for elastic scalability and
next-generation, high performance cloud
environments
In 2005, IBM launched its distributed caching technology as a part of IBM WebSphere Extended Deployment (XD). That technology became known as IBM WebSphere eXtreme Scale software and started the WebSphere caching product family. WebSphere eXtreme Scale operates as an in-memory grid that dynamically processes, partitions, replicates and manages application data and business logic across hundreds of servers. It provides transactional integrity and transparent failover to ensure high availability, high reliability and consistent response times. It is an essential distributed caching platform for elastic scalability and next-generation cloud environments. WebSphere eXtreme Scale provides the technology to enhance business applications, including web commerce, supply chain, financial trading and even on-line gaming, to form new, innovative classes of business applications by extending the data-caching concept with advanced features. In 2010, IBM introduced the WebSphere DataPower XC10 appliance (see Figure 1). IBM WebSphere DataPower XC10 is a purpose-built, easy-to-use appliance designed for simplified deployment and hardened security at the caching tier of your enterprise application infrastructure.
for rapid, “drop-in” use in conjunction with WebSphere Application Server and other WebSphere family products where it can deliver a cost-effective, distributed caching solu-tion in support of data-oriented distributed caching scenarios. Let’s take a more in-depth look at the scenarios and configu-rations addressed by this family of products:
● Simple Data and Database Cache: Applications can access data structures to improve performance and through-put using the WebSphere eXtreme Scale configuration as a local cache. Caching the data in memory attacks what is often the largest part of a transaction’s processing time— waiting for the data to arrive. WebSphere eXtreme Scale enhances this traditional data-caching scenario even further by offering failover capability. The DataPower XC10 appli-ance stores data on its Solid-state Disk (SSD).
● Client/Grid with Near Cache (software only): Having the data in memory is a proven step to higher performance. When dealing with large volumes of data, applications can perform even better. A Java™ Virtual Machine (JVM) can have a local eXtreme Scale grid, which sits in front of a remote grid serving as a “near cache” for a subset of the data, allowing a client to leverage a very large remote cache to offload back-end processing or to speed access to cached results. The near cache is in the same JVM as the applica-tion and provides local, in-process access to data. It contains
● Side cache: When applications need information that is used frequently but does not change often, like a user’s pro-file, a side cache offers significant performance benefits. WebSphere eXtreme Scale can be used as a side cache to store objects that have been retrieved from the back end. The application checks the side cache first to see if it con-tains a record. If there’s a cache miss, then the data is retrieved from the backend and inserted into the cache. ● Real-Time Data and Event Mining (software only):
When working with real-time data flows, the first challenge is filtering and organizing the data so the applications can use it. A partitioned eXtreme Scale configuration can sub-scribe to events and apply them to partitioned data thus supporting near-linear scalability and deterministic latency for these applications.
● Map/Reduce support (software only): WebSphere eXtreme Scale clients can invoke agents that run against massive amounts of data on multiple nodes in parallel. Clients can then aggregate and further process the results stored in the grid by the nodes. WebSphere eXtreme Scale caches can potentially span thousands of JVMs and support extremely large datasets. In addition, WebSphere eXtreme Scale includes efficient new algorithms to allow in-memory caches to grow elastically as the number of available JVMs or physical machines changes. Traditional distributed cache products use Map APIs as their primary programming model. WebSphere eXtreme Scale offers this and also allows graphs of objects to be easily pushed to the cache. WebSphere eXtreme Scale allows simple Java objects to be annotated and uses a simpler API to transparently allow these graphs to be fetched from the grid as well as to push any changes made by the application back to the grid. This significantly simplifies programming compared with the older JCache or Map-based APIs, improving productivity of the application developers by allowing them to focus on the core business logic.
Along with the above scenarios, the WebSphere caching fam-ily of products provides a comprehensive elastic persistent store solution for business applications while offering:
High performance:
● In-Memory management: Includes in-memory manage-ment capabilities with the capacity to support terabytes of data spread across thousands of servers. This allows your business application to seamlessly scale with the data growth. Note: The appliance has a large cache (stored on its SSD) as previously stated.
● Fast application performance: Data is stored in memory so applications have extremely fast (millisecond) access to data. Being able to access data in memory means faster applications and less workload for the supporting database servers. Note: The DataPower XC10 appliance stores data on its SSD to enable fast application performance. ● Flexible definition of data location: Ability to specify
“zones” where data should reside, either within a server system, data center, or geography and have the system transparently manage data placement to those zones. A zone can be a chassis, a building or a different geography. Synchronous replicas reside in the same zone as primary replicas with asynchronous replicas residing in different zones. This offers more granularity of shard placement, providing more fault tolerance. This allows your data store to support the availability constraints unique to your busi-ness. Note: The DataPower XC10 appliance offers simple zones support.
Ease of use:
● Dynamic cache support for enterprise applications: As a cache provider for the WebSphere Application Server dynamic cache service, the WebSphere caching family pro-vides a consistent distributed “drop-in” cache for enterprise applications. This brings a higher quality of service, near-linear scalability and high availability to bear on a broad variety of business applications utilizing, or interested in utilizing, the WebSphere Application Server dynamic cache service with minimal invasive changes.
● Reporting and monitoring: WebSphere eXtreme Scale includes implementations of metric access adapters to improve integration with IBM Tivoli® Monitoring or Hyperic HQ, enabling comprehensive insight into the operational behavior of business solutions. DataPower XC10 includes a built-in console for ease of management and metric tracking.
● Ease of use for common transaction tasks: WebSphere eXtreme Scale handles many of the common retry and exception logic tasks within the grid middleware. The request timeout for clients removes the burden from devel-opers for boilerplate retry logic for most map interaction operations. Most retry conditions are handled automatically, enabling developers to focus on the business logic aspects of application development.
● Accelerated Time to Value (appliance specific): The WebSphere DataPower XC10 Appliance reduces the time necessary for install, setup and configuration through out-of-the-box, “drop-in” use for HTTP Session Replication and brings more efficiency and higher quality of service to the WebSphere Application Server dynamic cache service. ● Simplified management and administration (appliance
specific): The WebSphere DataPower XC10 Appliance offers a built-in, simplified administration and monitoring console to enable efficient setup, configuration and man-agement of the appliance and transaction load within your datacenter.
● Simplified monitoring of the runtime/health of the appliance (appliance specific): The WebSphere DataPower XC10 Appliance includes status widgets to report key metrics pertaining to your transaction load and memory. Two examples of the reported metrics are mem-ory/disk usage and average response time.
Key feature functions:
● Enhanced XTP capabilities:
– Building multitenant applications has been greatly sim-plified. Map templates allow applications to create new maps on demand, avoiding the need to use application discriminators in the keys or to create extra maps that may never be used.
– (Software only) Implementation of byte array maps reduces the memory footprint of the grid cache, result-ing in lower memory costs and better performance. – (Software only) The indexing supports multiple
attrib-utes. This feature can simplify index usage when query-ing multiple attributes and reduce the overhead of having multiple indexes defined. Query has also been optimized to take advantage of composite indexes. – (Software only) Operational flexibility includes the
abil-ity to snapshot maps, which allows applications to easily generate snapshots for reporting purposes and with replica completion for incomplete replicas based on the last replica checkpoint and synchronous replication. – (Software only) With support for WebSphere Real Time,
● Write-through caching (software only): Write-through caching immediately propagates all changes from the in-memory cache to the back-end database as part of the trans-action. This method results in longer response times but guarantees all changes are persisted to the back-end data-base. This approach also provides synchronization between the cache and back end. This is valuable in situations where the data must be hardened to the back-end data store for the transactions to be considered complete.
● Write-behind caching (software only): Write-behind caching batches data updates, sending them to the backend data store at a configured interval. This dramatically improves transaction response times as they no longer need to deal with the back-end data store in a synchronous fash-ion. This approach reduces load on the data store and shields the application from back-end outages. The grid will hold the updates in memory in a fault-tolerant manner until the back end comes back online.
● JPA loaders (software only): Loaders are needed to read and write data from the data store when using WebSphere caching family product as an in-memory cache. Starting with WebSphere eXtreme Scale V 6.1.0.3, there are two built-in loaders that interact with JPA providers to map relational data to the eXtreme Scale maps, the JP Loader and JPAEntityLoader. The JPALoader is used for caches that store Plain Old Java Objects (POJOs) and the JPAEntityLoader is used for caches that store eXtreme Scale entities. These loaders reduce the burden on the application programmer.
● HTTP session replication: A servlet agent is provided that easily plugs into an existing application for scalable, fault-tolerant session management. Session data may be stored locally on the application JVM or remotely on the grid. Remote grid offers data replication for fault tolerance. Multiple data center support is available and replication between WebSphere Application Server cells is also sup-ported. There is a choice of synchronous or asynchronous replication depending on how rapidly the cached data changes offering a trade-off between zero data loss and exe-cution efficiency.
● Key-based replication (v7.1 software only): Now, WebSphere eXtreme scale supports key-based replication scenarios, which are also referred to as multimaster or Availability Partition (AP) replication scenarios. ● Non-Java client support—C/C++, .NET (software
only): WebSphere eXtreme Scale allows native, non-Java client access to its caching features via REST APIs, implemented on ADO.NET Data Services V1 specification. ● Improved reporting and monitoring (v7.1 software and
appliance): WebSphere eXtreme Scale is instrumented to work with DB2® performance monitoring tools. Its statis-tics are enhanced to enable querying to better work with other monitoring consoles, and a native administrative con-sole is now included.
● Data Grid Optimizations: WebSphere eXtreme Scale now has a smaller footprint of data in the grid. This results in less memory costs (lower TCO) and better performance. Additionally, you will enjoy a reduction in grid infrastruc-ture overhead via footprint and path length.
● High availability and fault tolerance: If a primary server fails, a replica is automatically promoted to be the primary. Multiple replicas can be supported. This can be done asyn-chronously or synasyn-chronously, and is transparent to the application. The result is a business application that is avail-able whenever the client needs it.
For ease of reference, the table below outlines the same key scenarios and features side by side for the two WebSphere caching family products:
WebSphere eXtreme Scale V 7.1 WebSphere DataPower XC10 Appliance
Optimized for: Developers/architects or businesses with a need to support transaction intense services cost-effectively by extending the functionality of existing
applications with fast and efficient transaction response times
Small to large businesses seeking a drop-in appli-cation scaling solution to quickly and efficiently
handle data-oriented scenarios such as HTTP session management, elastic dynamic caching and
web-side cache
Scenarios:
Simple data and database cache ●
Client grid with near cache ● 䊊
Side cache ●
Real-time data and event mining ● 䊊
Map reduce support ● 䊊
Appliance Enhanced Value (for 3 supported runtime scenarios):
● Elastic dynamic caching ● Session management ● Web-side cache
Large cache capacity per appliance n/a ●
Out-of-the-box scenario support n/a ●
Native appliance management and administration
n/a ●
Simplified monitoring of the run time/health of the appliance
n/a ●
Key Functional Offers: High Performance
In-memory/SSD management ●
Fast application performance ●
Flexible definition and data location ● 䊊
Ease of Use
Heterogeneous environment support ●
Dynamic cache support for enterprise WebSphere Application Server applications
●
Reporting and monitoring ●
Write-behind caching
JPA loaders (OpenJPA L2 cache support) HTTP session replication
Key-based replication (Asynchronous replication)
Non-Java client support—C/C++, .NET (ADO.Net data service/REST support) Improved reporting and monitoring Enhanced foundational capabilities High availability and fault tolerance Near-linear scalability
WebSphere eXtreme Scale V 7.1
● ● ● ● ● ● 䊋 ● ●
WebSphere DataPower XC10 Appliance
䊊 䊊 䊊 䊊 ● Other Comparisons
Distributed object cache across multiple servers in a LAN
Distributed object cache across multiple servers in a WAN
Advanced placement rules for primary/replica
Uses TCP/IP
Easy POJO programming Embeds in existing platform Spring Support
Tight integration with Java EE Application Servers
Please Recycle IBM Business Partner, or visit the following websites:
● ibm.com/software/webservers/appserv/extremescale ● ibm.com/software/webservers/appserv/xc10
Additionally, financing solutions from IBM Global Financing can enable effective cash management, protection from tech-nology obsolescence, improved total cost of ownership and return on investment. Also, our Global Asset Recovery Services help address environmental concerns with new, more energy-efficient solutions. For more information on IBM Global Financing, visit: ibm.com/financing
©Copyright IBM Corporation 2010 IBM Software Group
Route 100
Somers, NY 10589 U.S.A.
Produced in the United States of America June 2010
All Rights Reserved
IBM, the IBM logo, ibm.com, SMART SOA and WebSphere are trademarks or registered trademarks of International Business Machines Corporation in the United States, other countries, or both. If these and other IBM trademarked terms are marked on their first occurrence in this information with a trademark symbol (® or ™), these symbols indicate U.S. registered or common law trademarks owned by IBM at the time this information was published. Such trademarks may also be registered or common law trademarks in other countries. A current list of IBM trademarks is available on the web at “Copyright and trademark information” at ibm.com/legal/copytrade.shtml.