Network Attached Processing Enabled by Azul Compute Appliances

What is Network Attached Processing?

Optimized for service-oriented applications, network attached processing delivers centralized compute and memory capacity to applications built on the Java™ platform in the same way network attached storage provides centralized storage capacity. Applications launch from traditional servers and seamlessly gain access to up to 768 processors cores and 320 GB of heap space.

Network attached processing delivers these key benefits:

• 300% productivity improvement without any change to in-place resources
• Better service levels with consistent, faster application response times
• 300% increase in asset utilization with lower power and cooling requirements
• 50% lower infrastructure cost and complexity

Enabled by Azul Compute Appliances

Network attached processing is enabled by Azul Compute Appliances. Installed on a Gigabit Ethernet subnet in your application tier, Java workloads are offloaded from traditional servers via the Java-compatible Azul Virtual Machine, included with the appliances. No code changes are required for the applications to offload to the Azul platform. Once offloaded, the application gains access to larger memory heaps and as many processor cores as the application needs. This access is controlled via the Azul Compute Pool Manager, which uses customer-set policies to govern resource allocation.

Each Azul Compute Appliance is a shared resource that multiple applications can simultaneously access.

Each application is fully isolated from each other according to strict Java security rules.  Memory isolation and protection is further augmented with hardware enabled Virtual Partitions for each running VM process. 

Applications from heterogeneous hosts running different versions of Java (1.4 or 1.5) can all tap into Compute Appliances at the same and support Linux, Solaris, HP-UX and AIX operating systems.

Grow Top Line Profitability through More Efficient Datacenters

Powered by Azul Compute Appliances, network attached processing delivers greater efficiency to the data center by allowing SLAs to be met more consistently to deliver richer services. The ROI is significantly faster than with traditional servers alone and total cost of ownership is 50 percent lower on average. Without disruption to existing in-place resources or application code, your key enterprise applications become 3x more productive.

Manage your datacenter more efficiently:

• Deliver consistently high service levels under unpredictable load
• Accelerate time to market for strategic business changes
• Lower data center costs dramatically
• Simplify application scalability
• Increase data center asset utilization

Scale "Difficult Applications”

In today’s always on world, application response times can be the difference between profit and loss. Poor response time is often caused by an inability to efficiently scale or enhance "difficult applications”. These applications are typically constrained by technical limitations such as how many transactions per second a given server can provide, memory heap size, garbage collection pauses or even thread locks. Network attached processing addresses all of these problems and lets applications scale well beyond the limitations of traditional servers, enabling increased business capabilities and services at a fraction of the cost of traditional computing models.

Reduce Infrastructure Total Cost of Ownership by 50%

Customer research shows that the Azul approach reduces infrastructure TCO by 50% on average. Network attached processing cuts operating costs through a dramatic reduction in the number of servers and OS instances that must be managed, and by improving power, cooling and space efficiency by an order of magnitude

Infrastructure TCO reductions are achieved by:

• Dramatically reducing number of servers and OS instances that need to be managed
• Reducing power and space requirements by an order of magnitude
• Increasing hardware utilization by 3x or more
• Delivering commodity-level hardware price / performance
• Reducing number of network and SAN connections

Agile Infrastructure with Increased Efficiency

Applications are typically deployed on dedicated clusters of servers (silos of capacity). This approach results in a rigid infrastructure that is slow to respond to changing business requirements, runs at low utilization levels, and is complex to manage. Network attached processing makes the infrastructure significantly more agile by deploying compute capacity as a centralized, shared service. The shared capacity is governed and allocated between all the applications based on business-drive policies, and optimized in real time to maximize service levels without requiring manual intervention.

Benefits of an agile infrastructure:

• Shared resources drive asset utilization
• Dynamic alignment of resources to shifting business requirements
• Consistently high service levels under unpredictable load
• Faster time to market for new services and features
• Application isolation and resource guarantees

Reduce power and cooling requirements by an order of magnitude

Power and cooling requirements put tremendous pressure on enterprise datacenters. Power consumption rises steadily as enterprises deploy enormous quantities of servers and each new server generation becomes more power-hungry than the previous one. Network attached processing relieves these pressures by shifting the bulk of the workload to more energy-efficient Azul Compute Appliances. Azul appliances provide the required throughput while consuming an order of magnitude less power than a cluster of commodity servers. This power and cooling savingsfrees datacenter capacity to host new applications and deliver new services to drive top line revenue growth.

Maximum capacity at minimum consumption:

• Up to 16 processors (768 processor cores) and 768 GB memory per appliance
• Only 500W of power
• 14U of Rack Space
• Shared infrastructure that does the work of 100 servers*

* Results proven with GigaSpaces software in a test environment

Reduce Infrastructure Total Cost of Ownership by 50%

Customer research shows that network attached processing reduces infrastructure TCO by 50% on average. Network attached processing enabled by Azul Compute Appliances cuts operating costs through a dramatic reduction in the number of servers and OS instances that need to be managed, and by improving power, cooling and space efficiency by an order of magnitude.

Infrastructure TCO reductions:

• Dramatically reduce number of servers and OS instances that need to be managed
• Reduce power and space requirements by an order of magnitude
• Increase hardware utilization by 3x or more
• Deliver commodity-level hardware price / performance
• Reduce number of network and SAN connections

Massive Memory Heaps, No GC pauses

Network attached processing eliminates memory heap limitations and delivers hardware-assisted Pauseless Garbage Collection (GC). With up to 320 GB heap sizes, applications are able to handle significantly higher workloads while delivering new, advanced functionality. And developers no longer need to spend valuable time optimizing and tuning applications and can instead focus on growing the business.

What can massive heaps with no garbage collection pauses do for you:

• Handle increased throughput with fewer application instances
• Ensure consistently fast response times
• Deliver enhanced application functionality
• Eliminate development time spent shrinking heap requirements
• Work with larger live data sets or cache more data

Unprecedented Application Scalability

Faced with scaling challenges, IT departments typically react by simply adding more servers, with a side effect of increasing complexity and application tuning. Network attached processing enabled by Azul Compute Appliances solve this problem by delivering massive compute and memory resources and allowing each application instance to tap into as much capacity as it needs the moment it needs it. Optimized to run Java-based applications, Azul Compute Appliances offer unique hardware-assisted features such as Optimistic Thread Concurrency and Pauseless Garbage Collection to eliminate typical scaling bottlenecks.

Increased application scalability enables you to:

• Increase throughput with fewer application instances
• Handle unpredictable workload spikes while meeting service levels
• Eliminate fine-grained memory locks
• Stop tuning the application to fit into small heaps

REALTime Performance Monitoring

Application profiling and performance monitoring tools allow developers to improve the behavior of their applications. The Azul REALTime Performance Monitor (RTPM) tool, part of the Azul Compute Appliance, provides developers with deep, instant performance information about threads and locks both at the Java application level and at the VM level. This unparalleled capability relies on unique information provided by the Azul VM. RTPM consumes minimal resource overhead and can be used in production environments to provide more actionable data from real-world application usage.

RTPM enables developers to:

• Identify application performance bottlenecks with deep, detailed and real-time information
• Monitor performance in production environments with very low resource overhead
• Eliminate profiling tool setup time with RTPM embedded in the Azul VM

Boost Developer Productivity and Accelerate Application Time to Market by 50%

Does your organization have an increasing backlog of new services and features? Do your developers spend the majority of their time maintaining and optimizing existing applications instead of developing new ones?

By developing applications for Azul Compute Appliances, developers boost productivity and reduce application time to market. In a typical deployment, development teams spend up to 50% of their time capacity planning and optimizing applications to run on clusters of small servers. With network attached processing, developers no longer need to perform ‘unnatural acts of Java’ to tune applications for small server footprints. Instead, they can concentrate on developing new services, and get them deployed 50% faster.

With network attached processing your developers can:

• Maintain a strong pace of innovation
• Generate faster return on investment for new applications
• Increase competitiveness by delivering unique and highly differentiated services
• Stop tuning and start creating