Where in St. Petersburg to test intellectual technology of flash-storage HPE Nimble

What is included into HPE Nimble?

The storage system of HPE Nimble integrates in itself flash-arrays and HPE InfoSight – the platform of analytics and management of lifecycle. In a model range of HPE Nimble there are both all-flash-arrays for the main loadings, and the hybrid flash-systems integrating the capacity of flash-carriers and profitability of hard drives. Hybrid flash-arrays of HPE Nimble can be used both for the main storage, and for backup and disaster recovery of data in a combination with all-flash.

Other integral HPE Nimble component – HPE InfoSight. This solution personifies absolutely a different approach to increase in data availability and allows to prevent many faults even outside DWH. There is it thanks to machine learning and predictive analytics on its basis.

How does machine learning and predictive analytics of HPE InfoSight work?

At first there is collecting of telemetric information from sensors in storage systems and servers and also in their environment (including networks and the virtualization platform). Then training using algorithms of mathematical statistics and neural networks for work with unstructured data for identification of patterns in the data arriving from sensors from all devices worldwide is provided. On the basis of observation and training possible faults then the recommendations aimed at prevention of malfunctions, performance improvement and optimization of use of resources are made are analyzed. And at last, actions for the automated solution of problems and permanent improvement of experience with all environment are carried out to DPC.

Thanks to machine learning and predictive analytics in HPE InfoSight levels of support of L1 and L2 are automated. Statistics shows that an overwhelming part (93%) of requests for support opens automatically, and from them 86% are solved before users learn about emergence of malfunctions.

HPE InfoSight allows to collect and analyze data from all stack of infrastructure. More than in half of cases of unavailability of data of the reason, as a rule, are not in storage system, and in other IT components. On the basis of collected data of global telemetry an analytical system makes recommendations about elimination of problems in the present and the future. Proceeding from data on work of an array and the amounts of data stored on it, HPE InfoSight provides forecasts of loading of a system and capacity, simplifying planning of upgrades on the volume of storage and/or controllers of DWH.

How data protection is provided?

HPE InfoSight provides multiple-factor data protection. The producer guarantees availability of 99.9999% with the automated HPE InfoSight support (these are no more than 30 seconds of idle time a year).

In terms of information security the product will not create vulnerability for user applications as it has no data access and to a program code. HPE InfoSight has access only to telemetry of an array and environment. And even the telemetry gathers and transferred on the protected protocols, and pro-active actions are carried out only from approval of the system administrator.

The following factor – fault tolerance on the disks Triple + Partity RAID. The storage system allows failure of three disks in one RAID group without data loss and without stopping access to them. If it happens, a system just isolates sectors with failures and also reserves sectors on SSD for fast recovery.

Besides, HPE Nimble controls integrity of data for read operations and record in an array and also protects them using instant pictures which are executed without influence on performance. This mechanism is built in write process and indexings of data. Together with local security of information in storage systems replication and also data encryption is used.

What else approaches and technologies are used by HPE Nimble?

For the effective organization of data storage on arrays several approaches are used. First, there is more formatted capacity after formatting and assignment of space under OS. Besides, checksums and reserve areas due to work with SSD are applied.

As the stored data are optimized:

1. Deduplication of data with the variable size of the block which is optimized under certain applications and works on the fly.
2. the Compression of data on the fly which is started after deduplication. In the course of work the storage system monitors changes of load of the controller and depending on it switches the used compression algorithms.
3. Elimination of zero blocks from a data stream.
4. Thin selection of volumes of data, zero-copy cloning of volumes.

At the expense of the optimized data storage repeated economy of space in comparison with the formatted capacity received after formatting of DWH is reached. The producer supports efficiency of these methods also with what for all all-flash-arrays gives a guarantee of HPE Store More: its essence that arrays of HPE Nimble all-flash will give more effective capacity on each terabyte, than any other competitive all-flash-array.

In addition, all-flash and hybrid flash-arrays of HPE Nimble are simple and convenient in installation, setup and management. For storage systems perhaps horizontal (on volume) and vertical (on performance) scaling, it is also possible to integrate them in the cluster containing up to four arrays with general management and capacity.

Estimate possibilities of the HPE Nimble platform for the company personally in a democentra of KOMPLIT.

Notes