Storage Architecture for Storage Class Memories
a storage class and memory technology, applied in the direction of input/output to record carriers, instruments, computing, etc., can solve the problems of reducing the write bandwidth available to applications trying to write data, the limited amount of media that can be put into an array, and the inability to write data, so as to reduce the read and write sequence times
Inactive Publication Date: 2019-02-14
VEXATA INC
View PDF1 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Benefits of technology
This solution significantly reduces latency and increases throughput by bypassing CPU bottlenecks, allowing for higher concurrent write and read transactions and scalable metadata management, while maintaining low latency and high availability.
Problems solved by technology
However, keeping the meta-data DRAM 120 also means the amount of media that can be put into an array is limited because the DRAM 120 associated with the CPU 130 has a maximum limit in terms of the amount of addressable memory locations that are available.
The CPU overhead to complete this replication operation significant reduces the write bandwidth available to applications trying to write data.
Erasure coding is demanding on CPUs, as the CPU is responsible for performing the computation pertaining to the erasure code as well as performing the cascaded I / Os to the various nodes that make up the storage cluster and significantly reduces the write bandwidth that is available to applications.
However, a system that does not provide high availability guarantees is never properly used to store critical data and is not considered a primary storage array.
There are two critical limitations of this approach: 1) Bandwidth bottlenecks with the CPU 130, and 2) the storage stack is typically a SCSI stack from the network->CPU and from the CPU->drives.
A big latency penalty occurs as the signaling must go through the storage stack twice.
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0027]Described herein are a number of different improvements to storage architectures, storage devices, and methods of using the same. It will be appreciated that in this description, reference is made to aspects that are pertinent to the understanding thereof, and that additional specific design and implementation details, which would be apparent in light of the descriptions provided, are not set forth, as to do so would obscure the significant features described herein.
[0028]It is also noted that in the descriptions herein, the processing unit (whether SPU or CPU as hereinafter described) are described as handling or performing certain operations. It is understood and intended by this vernacular that a software application exists that contains instructions therein, which instructions are then executed by the processing unit, as is known.
[0029]In one aspect is described an improved storage architecture that eliminates data path handling from the main control CPU as conventionally ...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
Described is an improved storage architecture. In a particular aspect an improved storage architecture with increased throughput to Ethernet storage modules due to elimination of data path handling from a main control CPU is set forth. Other method and apparatus are described therein, including a scalable Ethernet storage module particularly suited for usage with the improved storage architecture described herein.
Description
[0001]This application is a continuation of U.S. patent application Ser. No. 14 / 562,110, filed Dec. 5, 2014, the contents of which are incorporated by reference herein.BACKGROUNDField of the Art[0002]Described are a method and apparatus for improved storage, and, particularly an architecture that allows for increased throughput, increased scalability, and reduced overhead.Brief Description of the Prior Art and Observed Deficiencies[0003]Existing storage architectures are designed with hard disk drive (“HDD”) (spinning disks) in mind. Storage traffic from the external world comes in to an application server. The application server, which is separate from the conventional storage array and connected to the traditional storage array system through a network, makes an I / O request that traverses the network and terminates in the storage array system through the target mode host bus adapter (“HBA”) and is first placed in CPU memory. An example of a conventional storage architecture 100 is...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & AuthorityApplications(United States)
IPC IPC(8): G06F3/06G06F13/00
CPCG06F13/00G06F3/0656G06F3/0613G06F3/065G06F3/0688G06F3/0689
InventorRAVINDRAN, VINODHALTEKAR, SATSHEELVADIVELU, RAMKUMARNAGAPUDI, VENKATESHVARANASI, SURYA P.HUSSAIN, ZAHID
OwnerVEXATA INC