2. Oracle 10g Database Storage Demystified Jeff Browning, O.C.P, R.H.C.A. Senior Manager Network Appliance, Inc. OracleWorld 2003 San Francisco

3.  A little history  The notion of storage networking  SAN and NAS – Current-technology SAN: FCP – Current-technology NAS: IP over GbE  RAID: The “packaging” of hard disks – RAID0 – RAID1 – RAID4 – RAID5 – Combinations of RAID levels  Emerging storage technologies – ATA RAID – Serial ATA (SATA) – iSCSI – NFS v. 4 (NFS RDMA)  Conclusion and wrap up Agenda

4. A Little History  IDE/ATA: The beginning  SCSI: A proliferation of standards – SCSI-1 – SCSI-2: The proliferation begins – SCSI-3: A new approach

5. In the Beginning There Was IDE/ATA  Introduced by IBM with the AT/PC in 1984  Supported a master/slave concept  Enhanced and adopted by Compaq in 1986 with the Deskpro 386 as the IDE interface – ATA and IDE are now interchangeable terms

6. What You Could Do with an IDE/ATA Device: Not Much  IDE/ATA was slow (4 MB/s to start)  It didn’t support many devices (usually 2 hard drives)  It wasn’t reliable  But it was, and remains, very, very cheap  It was never used widely for databases

7. SCSI: A Proliferation of Standards  Invented by Alan Shugart (founder of Seagate) in 1979  Adopted as an ANSI standard in 1986  First version was referred to as SCSI-1

8. What You Could Do with a SCSI-1 Device: A Bit More  SCSI-1 was still pretty slow (5 MB/s)  It supported 7 peripheral devices  It was more reliable than IDE/ATA  It was also more expensive  This was the first choice for Sun, HP and other open systems vendors and, notably, the Macintosh

9. SCSI-2: The Proliferation Begins  Fast SCSI: Higher transfer speed (10 MB/s or higher)  Wide SCSI: Width of the bus was increased from 16 to 32 bits  More devices per bus (from 7 to 15)  Other improvements – Improved cables and connectors – Improved signaling – Active termination

10. SCSI-3: A New Approach  With SCSI-3 the approach changed – Cabling and connection layer no longer defined in the basic spec  So-called “interconnect” or “physical layer” standards  SCSI-3 basic spec only defines a command set and a communication protocol

11. SCSI-3: The Physical Layer Standards Serial Bus SCSIThis is the form of SCSI-3 found in many hosts today Serial Storage Architecture (SSA) Used by IBM on its larger systems; not common Fibre Channel Protocol (FCP) Defines a standard for SCSI-3 traffic over Fibre Channel networks; by far the most popular form of SCSI-3 today for databases iSCSIEmerging standard for SCSI-3 traffic over IP networks

12. The Notion of Storage Networking  SCSI provided a way to attach disks to a host  The need for sharing of disk and tape backup resources led to the idea of “shared SCSI”

13. Storage Networking for Applications  Certain applications required shared disk  Shared SCSI evolved as a way to solve this problem

14. Storage Networking Evolves  Storage networking evolved along two paths – SAN: With FCP being the dominant protocol – NAS: With Gigabit Ethernet (GbE) NAS became a viable alternative to FCP for many applications  The next section discusses the tradeoffs between these approaches

15. SAN and NAS  Storage Area Networks (SAN) take the approach of making SCSI sharable  Network Attached Storage (NAS) uses existing file sharing protocols to connect databases to storage  Both approaches have their place: They are different

16. Fibre Channel Emerges as Dominant SAN  Fibre Channel was designed as a SAN protocol  It was adopted as an ANSI standard in 1994  It has emerged as the de facto standard for creating a SAN

17. Typical Fibre Channel SAN

18. Fibre Channel SAN Tradeoffs  Advantages – Bandwidth is good: 2 Gb FC is now common – Host CPU cost per I/O is comparable to SCSI – Latency is low and performance is good – Scalability is good  Disadvantages – More expensive than comparable IP network – Interoperability is poor but improving – Highly complex to setup and administer – Difficult to share disk capacity

19. NAS Emerges as Alternative to SAN  NFS was created by Sun in in the early 1980s  Version 1 of NFS was widely regarded as inappropriate as a file sharing protocol for databases  Version 2 improved enough that Oracle certified NFS for Oracle datafiles in 1997  Version 3 builds upon those improvements  Version 4 is emerging (more on this later)

20. Typical IP/GbE NAS

21. IP/GbE NAS Tradeoffs  Advantages – Bandwidth is pretty good using GbE – Switches/NICs are very inexpensive compared to FC switches/HBAs – Simple and easy to setup and administer – Interoperability is excellent – Disk capacity can be easily shared – even across platforms  Disadvantages – Host CPU cost may be higher than FC, depending on load, but not if the load is spindle-bound (NFS v. 4 fixes this in spades) – CPU Scalability (in the sense of CPU count) can be lower than FC (again NFS v. 4 addresses this)

22. SAN vs. NAS Suitability  SAN – Suitable for high-end environments where latency, performance, or CPU cost per I/O are critical – Required by some applications where NAS is not supported  NAS – Suitable for low- or mid-end environments where performance or CPU cost is less important than $$ cost – Also suitable for some high-end environments where CPU is compute intensive, not I/O intensive  SAN and NAS are converging

23. RAID: Redundant Array of Inexpensive Disks  The problem: – Disks are fragile; they fail – Data is precious and must be protected – Tape or disk backup is too slow or too expensive  RAID provides a way to combine disks together with redundancy so that a single disk failure will not lose data  Hot spares and auto-promotion make this a viable long-term solution  Software RAID vs. hardware RAID

24. RAID and Its Variants RAID0Simple striping; not truly RAID RAID1Disk-to-disk mirroring RAID4Striping with a parity disk RAID5Striping with striped parity RAID1+0 RAID0+1 RAID5+1 Etc. Combinations of RAID protection; can get complex

25. RAID0: Striping

26. RAID0 Tradeoffs  Advantages: – Fastest type of RAID; leverages disks well – No disk overhead  Disadvantage: – A single disk loss is critical  Suitability – Any environment where performance is important, and you do not care about the data, e.g. Datamarts

27. RAID1: Simple Mirroring

29. RAID1 Tradeoffs  Advantages: – Read capacity is higher than single disk (but lower than striping) – Very fault tolerant; all data is mirrored  Disadvantage: – Single disk capacity for writes – Two write per I/O penalty – Doubles disk cost  Suitability: – Very commonly used for online redo logs

30. RAID0+1: Striping with Mirroring

32. RAID1+0: Mirroring with Striping

34. RAID0+1/RAID1+0 Tradeoffs  Advantages: – Read capacity is high; multiple disks are leveraged – Very fault tolerant; all data is mirrored  Disadvantage: – Two write per I/O penalty – Doubles disk cost  Suitability: – Very common for storing Oracle datafiles where redundancy is highly valued

35. RAID4: Striping with Parity Disk

37. RAID4 Tradeoffs  Advantages: – Read Capacity is high; multiple disks are leveraged – Low RAID overhead; almost as good as RAID 0 – RAID protection exists  Disadvantage: – Two disks cannot be lost – Parity disk can become a bottleneck (some vendors avoid this issue with buffering, in which case performance is similar to RAID 1)  Suitability: – Very common for storing Oracle datafiles where redundancy is needed, and the cost of RAID0+1/RAID1+0 is too high

38. RAID5: Striping with Striped Parity

40. RAID5 Tradeoffs  Advantages: – Read Capacity is high; multiple disks are leveraged – Low RAID overhead; almost as good as RAID 0 – RAID protection exists  Disadvantage: – Two disks cannot be lost – Slowest RAID; CPU cost of parity striping is high  Suitability: – Very common for storing Oracle datafiles where redundancy is needed, performance is not critical, and the cost of RAID0+1/RAID1+0 is too high

41. Emerging Storage Technologies  ATA RAID  Serial ATA (SATA)  iSCSI  NFS v. 4 (NFS RDMA)

42. ATA RAID  A repackaging of cheap ATA/IDE disks  Used as a tape backup substitute  Archive storage is on-line and accessible  Faster than tape  Almost as cheap as tape, or even cheaper if compression is used

43. Serial ATA  An updating of the ATA/IDE spec to current technology  Intel and Dell  Targeted for desktops and next generation storage appliances  Could become a serious competitor to FCP and serial bus SCSI

44. iSCSI  Implements SCSI-3 protocol over IP networks  Intel is a leader  Software initiators exist for Windows and Linux  HP-UX and AIX initiators are in public beta  Targets are available from a variety of vendors  Presently immature, but will become viable competitor to FCP – Key is TOE HBAs on both target and initiator  Effectively offloads host/target CPU from IP traffic – Cost per port for switches and HBAs is vastly cheaper than FCP – If performance becomes comparable, FCP could be toast

45. Typical iSCSI SAN

46. NFS v. 4 (NFS RDMA)  Basically, a rewrite of NFS  Focused on “local sharing” i.e., database customers and the like, who need to share data across a small, focused network with very good performance  Supports Read Direct Memory Access, a very high performance, low latency I/O protocol  Supports Infiniband as an I/O interface  Leaders are Network Appliance and Sun  Will provide a transparent performance upgrade path for NFS database customers

47. Agenda  A little history  The notion of storage networking  SAN and NAS – Current-technology SAN: FCP – Current-technology NAS: IP over GbE  RAID: The “packaging” of hard disks – RAID0 – RAID1 – RAID4 – RAID5 – Combinations of RAID levels  Emerging storage technologies – ATA RAID – Serial ATA (SATA) – iSCSI – NFS v. 4 (NFS RDMA)  Conclusion and wrap up

48. Wrap Up