Elastic Scaling Servers
Thanks to automation technology, we are generating data faster than ever, but most organizations still need to derive insights from their data. Many modern servers can't manage (capture, store, and analyze) the flood of data, causing organizations to fall behind. Many organizations can't cope with "Big Data" (defined as any data that is beyond the recipient's ability to manage) despite owning large–scale systems. The solution lies in Scaling Servers™. These ultra–powerful systems can capture and store vast amounts of data in real time, and process it quickly and efficiently for visualization on smaller platforms. Businesses that haven't discovered these new systems may not be performing the routine analytics they need to remain competitive.
What Are Scalable Servers?
When a server isn't pushed beyond its data processing limit, it will perform efficiently, but when the data exceeds the server's limits, efficiency is lost. Scaling is a method of dividing the data among multiple servers that can then process it in parallel, keeping each server within its efficiency range and reducing overall processing time. Some scaling systems use a very complicated algorithm that adds each processing node "programmatically." These systems tend to be complicated and costly, typically offering only a few nodes, and they don't scale without additional third–party software.
What Is Elastic Scaling?
Scientel's Scaling Servers employ a Linear Scaling Technology. That means our systems come "out of the box" ready to expand and scale to thousands of nodes. This allows faster adjustment of the number of processors/nodes, up or down, to match the speed of the system to the load (volume and variety of data). Scientel also supports the development of single applications that can process different data sets on different nodes based on the analytic requirements that are demanded by users while still in parallel mode.
The Scientel Model ESS 3400 High–Level Elastic Scaling Server™ (ES2) system represents the true state–of–the–art in HPC database computing by achieving several million transactions per minute. This performance is achieved with Scientel's GENSONIX Enterprise NewSQL Super DB, which can scale to 3456 DB nodes on the 3400 model. GENSONIX Enterprise DB can perform the equivalent of a join of BILLIONS of rows with TRILLIONS of transactions!
The ESS 3400 system is designed to reach over 1 Petabyte in storage and 166 TB of main memory for computing intense Big Data problems. It incorporates 1728 processor cores which can support 3456 DB nodes when running the GENSONIX Super DB. The ESS 3400 system is configured in a 4 full rack configuration which utilizes high–speed InfiniBand interconnect for high speed operation. InfiniBand configuration on this system supports data speeds of up to 100GB per second.
GENSONIX incorporates multiple data store schemes, such as Relational tables and Document stores which helps with data storage efficiently. GENSONIX has built–in intrinsic functions and array processing capabilities along with recursive DB sequencing to allow processing of large numbers of data rows at higher speeds.
Elastic Scaling Server: ESS 3400
- 108 Cluster nodes — 1728 processor cores, 166 TB of memory, 1 PB of storage
- Multi-million Transactions per minute
- Scales to 3456 compute nodes
- Gensonix NewSQL DB
- Distributed Storage Architecture(DSA)
- Expandable to multi-Petabytes of storage
- High-Performance, Ultra-Parallel Processing
- Modular rack configuration
Large Data Warehouse Appliances
The Scientel® LDWA Series is optimized to run GENSONIX Enterprise NoSQL DBMS to provide management and maintenance of SUPERDATABASES which range into TRILLIONS of transaction records.
When bundled with GENSONIX, our LDWAs can support transaction DBs of MULTI–TRILLION records, incorporating several TRILLIONS of data elements organized in column structures, network structures, etc., with real–time query capabilities.
Running on any of our LDWA systems, GENSONIX Enterprise DB can perform the equivalent of a join of BILLIONS of rows with TRILLIONS of transactions!
All of our LDWA Platforms share the following features:
- Achieving Multi–Million Transactions per second
- Scales to very large # of compute nodes and CPU cores, e.g. 1000s
- Multi–terabyte+ scalability
- High Uptime
- High reliability
- Expandable to multi–Petabytes of storage
- High–performance, massively–parallel architecture
- Modular rack configuration
Entry Level System: LDWA 1100
- 11 nodes
- 96 Intel processor cores
- 128 GB of memory
- 100 TB of storage
Base Level System: LDWA 2200
- 22 nodes
- 176 Intel processor cores
- 264 GB of memory
- 160 TB of storage
Mid Level System: LDWA 4400
- 44 nodes
- 328 Intel processor cores
- 544 GB of memory
- 480 TB of storage
Mid Level System: LDWA 8800
- 88 nodes
- 704 Intel processor cores
- 1TB of memory
- 1PB of storage