Advances in Non- Vola0le Storage Technologies Tom Coughlin Coughlin Associates Ed Grochowski, Computer Memory/ Storage Consultant

Transcription

1 Advances in Non- Vola0le Storage Technologies Tom Coughlin Coughlin Associates Ed Grochowski, Computer Memory/ Storage Consultant 2015 Coughlin Associates 1

2 Outline The Shape of Things to Come In Search of New Memories Intel/Micron s 3D XPoint Technology The Memory is the Computer Conclusions 2015 Coughlin Associates 2

3 The Shape of Things to Come 2015 Coughlin Associates 3

4 IOPS Required Respondents 40% 30% 20% 10% 79% between 10K and 1M IOPS 0% K 10K 100K 1M 10M IOPS From the 2014 How Many IOPS Do You Really Need Report, Coughlin and Handy, Coughlin Associates 4

5 Maximum Latency Requirement 40% Respondents 30% 20% 10% 36% at 10ms latency 0% <10ns 100ns 10µs 1ms 100ms >1 sec Latency From the 2014 How Many IOPS Do You Really Need Report, Coughlin and Handy, Coughlin Associates 5

6 Latencies Separate Computer Memory from Storage Doug Voight, HP, 2015 FMS Non- Uniform Memory Access 2015 Coughlin Associates 6

7 Touch rate versus response Ume indicaung various types of uses 2015 Coughlin Associates 7

8 Digital storage technologies regions overlaid on the Touch Rate/Response Time chart Memory 2015 Coughlin Associates 8

9 Ent. SSDs, perf. HDDs and capacity HDDs 2015 Coughlin Associates 9

10 Storage Latency with Current and Future NV Solid State NV Technologies (from NVMP talk at SNIA Winter Symposium, 2014) SoOware 2015 Coughlin Associates 10!

11 ImplicaUons of Persistent Memory NV memory retains its data even when power off thus instant recovery of state before power down is now possible NV memory with the right SW changes can provide much be_er latencies than current systems and SW Non- volaule memory will save power since refreshes not needed Persistent memory creates new opportuniues to share that memory between different computers or computer chips using Remote Direct Memory Access (RDMA) Embedded NVM technology can lead to logic- in- memory architecture for future SoC this can lead to new distributed computer architectures 2015 Coughlin Associates 11

12 Challenges of Persistent Memory If the memory isn t cleared by rebooung a system, then reboots to recover funcuonality with corrupted data won t work There will need to be a special reboot and clear funcuon that erases and recovers data in memory Or, and this may be an even be_er soluuon, we need to built devices and soeware that are self- monitoring and self correcung so we don t need to reboot in order to retain funcuonality 2015 Coughlin Associates 12

13 In Search of New Memories 2015 Coughlin Associates 13

14 Memory and Storage Today 2015 Emerging NVM and Their Manufacture Report, Coughlin Associates 2015 Coughlin Associates 14

15 Hard Disk Drives Boilerplate 2015 Coughlin Associates 15

16 HDD Areal Density Growth Recent Developments: SMR, HAMR, He- filled HDDs, MAMR with 3D Recording, 2015 Coughlin Associates 16

17 Actual Quarterly Announced HDD Product AD Charts % CAGR Gbits per square inch % CAGR 60% CAGR 40% CAGR 20% CAGR HDD Product Coughlin Associates, Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Coughlin Associates 17

18 Flash Memory 2015 Coughlin Associates 18

19 NAND Technology Roadmap J. Choe FMS Coughlin Associates 19

20 Annual HDD/SSD Storage Shipments Annual Shipped Storage Capacity (Exabytes) 1, , HDD SSD ~10 X Coughlin Associates 20

21 Comparisons of Flash and HDDs Mike Workman Keynote, 2015 FMS 2015 Coughlin Associates 21

22 Emerging Memory Technology NVM will save power Persistent memory enables memory sharing (RDMA) Embedded NVM technology can lead to logic- in- memory architecture The emerging NVM market could exceed $2 B by Emerging NVM Report and Their Manufacture, Coughlin Associates 2015 Coughlin Associates 22

23 FeRAM! FeRAM products shipping for many years for niche applicauons, hard to scale density 2015 Coughlin Associates 23

24 Phase Change Memory (PCM)! Several industry experts think that Micron/Intel s 3D XPoint Technology is a crossbar phase change memory 2015 Coughlin Associates 24

25 TYPES OF RRAM TECHNOLOGY Sources/ Crossbar, Tezzaron Semi. Oxide Conduction Via Oxygen Vacancies Metal Filament Formation (CBRAM) 2015 Coughlin Associates 25

26 STT MRAM Replacing DRAM and SRAM Over 50 M MRAM chips shipped by Everspin 2015 Coughlin Associates 26!

27 NVM Progress (Amigo Tsutsui, Sony at the 2015 FMS) 2015 Coughlin Associates 27

28 Memory Selector Device OpUons An Chen, 2014 AVS TFUG Seminar 2015 Coughlin Associates 28

29 2015 Coughlin Associates 29

30 Annual Memory Capacity Shipments (2014 Emerging NVM Report and Their Manufacture, Coughlin Associates) 2015 Coughlin Associates 30

31 3D XPoint Technology 2015 Coughlin Associates 31

32 Comparison of Memory Technologies Slide from 2015 Storage Visions Conference NV- Memory Normalized Log Scale Capacity 1MB = 1unit Power/GB 10mW = 1unit Cost/GB 10c = 1unit SRAM DRAM NAND DISK NAND DISK Next- Gen Storage DRAM SRAM DISK NAND DRAM SRAM Latency 1ns = 1unit SRAM DRAM NAND DISK 1000X Next- Gen NV- Mem Next Gen Memory must fill the DRAM & Flash latency gap 2015 Coughlin Associates 32 32

33 3D XPoint Memory 2015 Coughlin Associates 33

34 3D XPoint R Comparisons to NAND Latency (ls) 10,000 Source: Storage Technology Group, Intel SSD NAND technology offers ~100X reduc0on In latency versus HDD I NVMe TM eliminates 20 μs of latency today 3D XPoint TM technology reduces NVM latency offering ~10x reduc0on in latency vs NAND SSD 0 Drive Latency ControllerLatency Soeware Latency HDD SSD NAND (i.e. SAS HBA) SSD NAND SSD +SAS/ +SAS/SATA +NVMe TM 3D Xpoint SATA +NVMe TM Technology claims are based on comparisons of latency, density and write cycling metrics amongst memory technologies recorded on published specificauons of in market 2015 memory Coughlin products Associates against internal Intel specificauons 34

35 3D Xpoint Memory 2015 Coughlin Associates 35

36 The Memory is the Computer 2015 Coughlin Associates 36

37 Spintronics From Tohoku University in Japan 2015 Coughlin Associates 37

38 Spin- Based CompuUng mlogic: All Spin Logic Device and Circuits for Future Electronics, Jimmy Zhu, et. Al, March Coughlin Associates 38

39 mcell Spin Logic Unit (J. Zhu, Carnegie Mellon Univ.) 2015 Coughlin Associates 39

40 STT Logic ApplicaUon (Reducing Dark Silicon) 2015 Coughlin Associates 40

41 What if Spin Transfer Replaced Charge Transfer (Current)? Spin transfer would not generate the heat that electrical currents generate This allow building powerful devices in small spaces Spin could be used for both processing and memory/storage 2015 Coughlin Associates 41

42 Conclusions Faster memory/storage is needed for modern applicauons The storage/memory landscape has more opuons than ever e.g. 3D XPoint Non- volaule memories will replace volaule memories NVM and processors want to come together Spin- based electronics are one example of a technology that could put memory and processing in one device 2015 Coughlin Associates 42

43 References 2014 How Many IOPS Do You Really Need Report, Coughlin and Handy, Coughlin Associates 2014 and 2015 Emerging NVM Report and Their Manufacture, Coughlin Associates 2015 Storage Visions Conference PresentaUons Touch Rate: A metric for analyzing storage system performance, Steven Heltzer and Tom Coughlin, 2015 InformaUon on all these documents are available at:h_p:// Coughlin Associates 43

44 Thanks 2015 Coughlin Associates 44