Sigmod08ssd slides

SIGMOD’08
SIGMOD’08

A Case for Flash Memory SSD in
Enterprise Database Applications
Sang-Won Lee Bongki Moon
Sungkyunkwan University University of Arizona

Chanik Park Jae-Myung Kim Sang-Woo Kim
Samsung Electronics Co., Ldt.
Ldt. Altibase Corp. Sungkyunkwan University

COMPUTER SCIENCE DEPARTMENT ACM SIGMOD, Vancouver Canada, June 2008 -1-

Magnetic Disk vs Flash SSD
Champion M-Tron Flash SSD
for 50 years 32GB 2.5 inch

New
challengers!
Seagate ST340016A
40GB,7200rpm
Samsung FlashSSD
32GB 1.8 inch

ACM SIGMOD, Vancouver Canada, June 2008 -2-

Trend in Market Today
• In mobile storage market
NAND flash memory wins over hard disk in mobile storage market
• PDA, MP3, mobile phone, digital camera, ...
Due to advantages in size, weight, shock resistance, power
consumption, noise …
• In personal computer market
Compete with hard disk in personal computer market
• 32GB Flash SSD: M-Tron, Samsung, SanDisk
Vendors launched new lines of personal computers with NAND flash
SSD replacing hard disk
• Apple, Samsung, and others


Market Trend in Prospect
• Price drops quickly
NAND flash is a lot cheaper than DRAM;
• ASP/MB of NAND < 1/3 of ASP/MB of DRAM as of 2007.
Still much more expensive than magnetic disk.
Annual drop in ASP/MB was about 60% in 2006.
Projected annual drop in ASP/MB is about 30-40% in next 5 years.
[Eli Harari@SanDisk, August 2007]

• Emerging Enterprise Market
NAND ASP was $10/GB in 2007. With 40% annual drop, it could be
$800/TB in 2012.
Not inconceivable to run a full database server on a computing
platform with TB-scale Flash SSD as secondary storage.


Technology Trend in Prospect
• NAND flash density increases faster than Moore’s law
Predicted twofold annual increase of NAND flash density until 2012
[Hwang, ProcIEEE’03]
Toshiba hopes for 512GB SSD by the end of 2009
• 30 nm chip-making process, Multi-level-cell (MLC)

• Bandwidth catches up
Samsung MCAQE32G8APP-0XA [2006]
• Sustained read 56 MB/sec, sustained write 32 MB/sec
Samsung, Mtron [Feb. 2008]
• Sustained read 100~120 MB/sec, sustained write 80~90 MB/sec
Intel-Micron’s 4-plane architecture + higher clock speed [Feb. 2008]
Samsung MLC-based 256GB SSD with SATA-II [May 2008]


Past Trend of Disk
• From 1983 to 2003 [Patterson, CACM 47(10) 2004]
Capacity increased about 2500 times (0.03 GB 73.4 GB)
Bandwidth improved 143.3 times (0.6 MB/s 86 MB/s)
Latency improved 8.5 times (48.3 ms 5.7 ms)
Year 1983 1990 1994 1998 2003
Product CDC Seagate Seagate Seagate Seagate
94145-36 ST41600 ST15150 ST39102 ST373453
Capacity 0.03 GB 1.4 GB 4.3 GB 9.1 GB 73.4 GB
RPM 3600 5400 7200 10000 15000
Bandwidth 0.6 4 9 24 86
(MB/sec)
Media 5.25 5.25 3.5 3.0 2.5
diameter
Latency 48.3 17.1 12.7 8.8 5.7
(msec)


Latency of Disk Lags
• Trend
In the time that bandwidth doubles, latency improves by
no more than a factor of 1.2 to 1.4.
• Latency improves by no more than square root of the
improvement in bandwidth.
The bandwidth-latency imbalance may be even more
evident in the future.
• The trouble is
Latency remains important for
• Interactive applications, database logging (or whenever I/O must
be done synchronously)
• What can NAND Flash Memory do for this?


Magnetic Disk vs NAND Flash
• Below is what the data sheets show
Sustained Transfer Rate Average Latency
Magnetic Disk 110 MB/sec 8.33 msec
NAND Flash SSD 56 MB/sec (read) 0.2 msec (read)
32 MB/sec (write) 0.4 msec (write)

Magnetic Disk : Seagate Barracuda 7200.10 ST3250310AS
NAND Flash SSD : Samsung MCAQE32G8APP-0XA drive with
K9WAG08U1A 16 Gbits SLC NAND chips

• Newer SSD products report much higher bandwidth for read and write


Characteristics of NAND Flash
• No mechanical latency
Flash memory is an electronic device without moving parts
Provides uniform random access speed without seek/rotational
latency
• Very low latency, independently of physical location of data
• Asymmetric read & write speed
Read speed is typically at least twice faster than write speed
• (E.g.) Samsung 16 Gbits SLC NAND chips: 80 µsec vs 200 µsec (2 KB)
• No in-place update
No data item or page can be updated in place before erasing it first.
• An erase unit (typically 128 KB) is much larger than a page (2 KB).
• (E.g.) Samsung 16 Gbits SLC NAND chips: 1.5 msec (128 KB)
Write (and erase) optimization is critical


Flash SSD for Databases?
• Immediate benefit for some DB operations
Reduce commit-time delay by fast logging
Reduce read time for multi-versioned data

• Still, many concerns to be addressed
Random scattered I/O is very common in OLTP
• Slow random writes by flash SSD can handle this?
Flash-aware design of DBMS?
Flash-friendly algorithms?
Flash-friendly implementation?


Transactional Log
SQL Queries

System Buffer Cache

Database Transaction Temporary Rollback

Table space (Redo) Log Table Space Segments


Commit-time Delay by Logging
• Write Ahead Log (WAL) T1 T2 … Tn
A committing transaction force-writes its SQL
log records
Buffer Log Buffer
Makes it hard to hide latency
With a separate disk for logging pi
• No seek delay, but …
• Half a revolution of spindle on average
• 4.2 msec (7200RPM), 2.0 msec (15k RPM) DB
With a Flash SSD: about 0.4 msec LOG

• Commit-time delay remains to be a significant overhead
Group-commit helps but the delay doesn’t go away altogether.
• How much commit-time delay?
On average, 8.1 msec (HDD) vs 1.3 msec (SDD) : 6-fold reduction
• TPC-B benchmark with 20 concurrent users.


HDD vs SSD for Logging
• With SSD for log
CPU better utilized
• By shortening commit-
time, and serving more
active transactions.
Leads to higher TPS

• Exaggerated by caching entire
DB in memory
• TPC-B to stress-test logging
Transaction commit rate
higher than TPC-C


Temporary Table Space
SQL Queries

System Buffer Cache




Temp Data and Query Time
• Query processing often generates temp data
Sorts, joins, index creation, etc.
Typically bulky, performed in foreground;
Direct impact on query processing time
• Typically stored in separate storage devices

• Ask the same question
What happens if SSD replaces HDD for
temporary table spaces?


External Sort: I/O Pattern
• External Sort algorithm runs in two phases
Sorted run generation
• Partitioned to chunks, sorted separately and, saved in sorted runs
• Read sequentially from table space, written sequentially into temp space
Merging sorted runs
• Read randomly from temp space, written sequentially into table space

• Dominant I/O patterns are sequential write followed by
random read
No-in-place-update limitation is avoided.
These are flash-friendly I/O patterns!!


External Sort: Performance
• HDD vs SSD as a medium for a temp table space
Sort a table of 2 M tuples (200 MB), with 2 MB buffer cache
• SSD is good at sequential write + random read
Almost an order of magnitude reduction in merge times


One Less Tuning Knob?
• Cluster sizes for Sorting?
• With a larger cluster
Disk bandwidth improves (by
hiding latency)
The amount of I/O may also
increase due to reduced fan-in
for merging sorted runs
• Flash SSD is
With low latency, not as sensitive
to the cluster size
2KB page was the best with the
max fan-in


Hash-Sort Duality a Myth?

• The I/O pattern of hashing is said to be
random write (for writing hash buckets) + sequential read
(for probing hash buckets)
As opposed to sort (sequential write + random read)

• If it’s the case, hashing is not flash-friendly.
Re-implement hashing to make it flash-friendly?
It appears already done by some vendors.
• The observed I/O pattern was quite similar to that of sort
(sequential write + random read)


Hash Join: Performance
• HDD vs SSD as a medium for a temp table space
Hash-join two tables of 2 M tuples (200 MB) each, with 2 MB buffer
cache
About 3-fold reduction in join time


Rollback Segments
SQL Queries

System Buffer Cache




MVCC Rollback Segments
• Multi-version Concurrency Control (MVCC)
Alternative to traditional Lock-based CC
Support read consistency and snapshot isolation
Oracle, PostgresSQL, Sybase, SQL Server 2005, MySQL
• Rollback Segments
When updating an object, its current value is recorded in
the rollback segment
To fetch the correct version of an object, check whether
it has been updated by other transactions
Each transaction is assigned to a rollback segment; old
images of data are written to the rollback segment
sequentially (in append-only fashion).


MVCC Write Pattern
• Write requests from TPC-C workload
Concurrent transactions generate multiple streams of append-only
traffic in parallel (apart by approximately 1 MB)
HDD moves disk arm very frequently
SSD has no negative effect from no in-place update limitation
800
Logical sector address (x1000)

700

600

500

400

300

200

100

0
0 100 200 300 400 500 600
Time (second)


MVCC Read Performance
C …
• To support MV read consistency,
T1 T0
B

A
100

50 A 100 A 200
I/O activities will increase
A long chain of old versions may have
T2
to be traversed for each access to a
frequently updated object
• Read requests are scattered
50
->

randomly
100 A:

Rollback segment
(2)

Old versions of an object may be
stored in several rollback segments
With SSD, 10-fold read time reduction
was not surprising
00
- 1
200 A:
>

Rollback segment
(1)


Database Table Space
SQL Queries

System Buffer Cache




Workload in Table Space
• TPC-C workload
Exhibit little locality and sequentiality
• Mix of small/medium/large read-write, read-only (join)
Highly skewed
• ~80% of accesses to 20% of tuples
• Write caching not as effective as read caching
Physical read/write ratio is much lower that logical
read/write ratio

• All bad news for flash memory SSD
Due to the No-in-place-update limitation
In-Page Logging (IPL) approach [SIGMOD’07]


Concluding Remarks
• Clear and present evidences that Flash memory SSD can co-
exist or even replace Magnetic Disk
Even now for logging, rollback segments and temp table spaces
Write optimization needed for database table spaces
• Flash-Aware DBMS Design is a must!
Flash-friendly algorithms, flash-friendly implementations
Need fresh new look at almost everything: Buffer management, B-
trees, Sorting and Hashing, Self-Tuning, File Systems, etc.


Sigmod08ssd slides

More Related Content

What's hot (20)

Viewers also liked (17)

Similar to Sigmod08ssd slides (20)

Recently uploaded (20)

Sigmod08ssd slides