Learn how HPE Synergy with HPE EPA supports a wide range of Big Data and analytics applications with the flexibility, elasticity, scalability, and optimized performance most modern applications require.
Are you looking for a Big Data infrastructure solution with composability? Do you want to leverage Composable Infrastructure for Real-Time Analytics with HPE Elastic Platform for Analytics (EPA)?
Data insights are the core of any digital transformation as they are the key to increased competitiveness, operational efficiency, and improved customer services. Many enterprises are increasingly using AI and analytics solutions to extract insights from the data generated from various sources like IoT systems, sensors, mobile devices, and customer interactions.
With data expected to double every year, next-gen big data and analytics applications are extremely demanding in terms of flexibility and scalability of compute and storage resources. The newer solutions offer advanced AI algorithms work on large amount of data across various industries for applications like connected cars, large scale genome studies, and health record analysis for cancer research.
A typical big data application has multiple phases starting from data ingestion (I/O intensive) to model training and validation (CPU/memory intensive). Each of these applications demands a flexible and optimized infrastructure for different phases. In addition, varied infrastructure requirements in terms of CPU, memory, and network are required for delivering the right performance, flexibility, cost, security, and manageability. For example, data scientists want the flexibility to run interactive and batch analysis with on-demand compute. At the same time data engineers/cluster administrators want to ensure the scalability and reliability of production workloads, and maintain governance and control over cluster resources and costs.
Top challenges with the current infrastructure
Complexity
- Accommodating heterogeneous deployment models such as VMs, containers etc. in a single enclosure
- Upgrading enterprise-grade tested firmware on all the nodes simultaneously
Lack of flexibility in provisioning infrastructure
- Selecting and scaling of storage, compute and network based on application/workload demands
- Lacking elasticity to release the resources when they are no longer needed by the applications
- Scaling independently from storage and compute nodes
Longer time to market
- Taking too long to get the system up and running for production
How HPE EPA with HPE Synergy addresses these challenges
Traditional converged or even hyperconverged infrastructure with fixed ratios of storage, networking, and compute resources do not provide enough agility to address these challenges to meet the needs of the big data applications.
The HPE Elastic Platform for Analytics (HPE EPA) along with HPE Synergy provides an ideal environment to address these challenges.
HPE EPA is designed with a modular architecture blueprint with disaggregated compute and storage blocks to deliver flexibility and scalability to support big data and analytics workloads. HPE EPA is composed of modular infrastructure building blocks, each of them optimized for storage density, heavy computation, networking or standard compute activities. Combining the right mix of building blocks, customers can build infrastructures optimized for their workloads.
Key highlights of the HPE EPA solution
Simplified composability
- HPE Synergy facilitates composability by assembling and re-assembling hardware resources of compute, storage and fabric on demand to accommodate diverse workloads. For example, when you require more disks to run interactive queries, you can attach more disks from the resource pool. Once you finish, they can be easily released back to the pool from the single management interface. This reduces underutilization of hardware resources and increases productivity. There might be a situation where if you have bought an HPE Synergy frame for the workloads like virtualization and now you want to run the analytics on the same hardware. This is possible because you can run multiple workloads in a single frame.
- HPE Synergy has multi-tenancy capability with easy resource management to mix and match a wide range of workloads and applications, for example it supports No-SQL database systems like Cassandra, in-memory applications like Spark, streaming applications like Spark streaming and also heterogeneous deployment models from bare metal to virtualization to micro services, including containers with Kubernetes, OpenShift, and Docker. This helps adapt to continuous changes in the application layer,
Easy management
- The combination of HPE EPA and HPE Synergy allows support for fine tuning the hardware and firmware while delivering maximum performance with operating efficiency for big data applications along with workload optimization. For example, if you want to deploy performance optimized compute nodes for data ingestion (like compute-hungry Kafka nodes), they can be easily configured within EPA blueprint and fast deployed within the HPE Synergy system. The OS can be deployed in about 3-to-5 minutes via a golden image and all the nodes can be deployed in parallel across the frame. The system update is simplified. After downloading a single set of tested firmware, the firmware upgrade can be done in parallel on all the nodes reducing about 70% of the operations time.
- With HPE Synergy unified API, both coding effort and time are reduced. A unified API provides full programmability, so you can provision bare metal infrastructure with a single line of code that abstracts every element of your Composable Infrastructure, increasing operational efficiency and rapid deployment of IT resources. This increases productivity and control across the data center by integrating and automating infrastructure operations and applications, eliminating time-consuming scripting of multiple low-level tools and interfaces.
Simplified infrastructure and cost savings
- Integrated power and cooling technologies in HPE Synergy Composable Infrastructure help reduce the number of power supplies. Improved cooling efficiency allows more air through the frame required for the compute tier of the HPE EPA platform. Innovative interconnect fabric solution simplifies the management connectivity, scales beyond single frame and saves costs with fewer switches. When we compare HPE Synergy to a traditional big data infrastructure with 12-node rackmount servers, HPE Synergy consumes: 28 fewer power cables; 32 fewer network cables; 20% less rack space
- HPE Synergy eliminates 95% of network sprawl compared to rackmount servers at the compute module of the HPE EPA, which converges traffic inside the frames connecting directly to the external LAN..
Reduced time to market
- HPE EPA flexibility, elasticity, security, and manageability can be further achieved using HPE Synergy Composable Infrastructure as compute block. Indeed, HPE Synergy provides fluid resource pools and software defined intelligence, allowing dynamic resources allocation and reorganization in minutes.
This illustration depicts a typical HPE EPA configuration on HPE Synergy architecture:
The advantage of HPE Synergy and HPE EPA architecture
The combination of the software defined infrastructure provided by HPE Synergy with the modular and optimized building block model of HPE EPA architecture reduces the time from concept to implementation from months to days. What’s more, the infrastructure is easily repurposed with the ability to reconfigure hardware quickly. This means that the configuration can adjust to the needs as they surface rather than having to predict what the needs will be in the future.
For more information on this architecture
Or contact your HPE representative.
Server Experts
Hewlett Packard Enterprise
twitter.com/servers
linkedin.com/showcase/hpe-servers-and-systems/
hpe.com/servers
ComputeExperts
Our team of Hewlett Packard Enterprise server experts helps you to dive deep into relevant infrastructure topics.