Focusing primarily on the technical aspect of databases on NetApp storage systems and provide commentary, views and solutions that are useful to the end-users.
On weekends, always I try to visit my parents in my hometown called Pirajuí.
Before arriving in Pirajuí and give a hug and big kiss on my parents and my family, I need to stop (a little bit) in a traditional Brazilian city very close my hometown.
The city of Bauru has an interesting history regarding a sandwich.
Bauru is a popular Brazilian sandwich. The traditional recipe calls for cheese (usually mozzarella) melted in a bain-marie, slices of roast beef, tomato and pickled cucumber in a French bread with the crumb (the soft inner part) removed.
The Bauru has a fairly well documented history. In 1934, a student at the "Faculdade de Direito do Largo de São Francisco", Casemiro Pinto Neto (known as Bauru for coming from that city), entered Ponto Chic, a traditional eatery and student hangout, and asked the cook to prepare a sandwich from his specifications. "Bauru's sandwich" was an immediate hit, and eventually became the best-selling dish at the place.
Unfortunately, many other eateries (like other storage company ^2), though, offer sandwiches named Bauru with different combinations of ingredients -- for instance, using sliced ham (EMC Snapshots) instead of roast beef (NetApp Snapshots) or sliced bread (EMC Control Station - based on Linux) instead of French bread (NetApp Unified Architecture).
Those are considered inferior to the "true" Bauru. The city of Bauru eventually named the traditional Bauru as the city's official sandwich, codifying the recipe in a municipal law and instituting an official certification program.
I love this history because we have a similar comparison in the storage world when some companies try to offer sandwiches (imitation) that is not the original Bauru (the perfect sandwich).
Please pay attention, only choose the original Bauru sandwich, where "Honestly, my sandwich is smarter..." - powered by Alex McDonald.
Another good example coming from Bauru region is the word: neto (the inventor of Bauru sandwich). Neto means:
Using mathematics is not hard to show that 1=2. This is called fallacy where an incorrect result arrived at by apparently correct, though actually specious reasoning.
The most common example of a mathematical fallacy is the "proof" that as follows. Let , then
The incorrect step is in which division by zero () is performed, which is not an allowed algebraic operation. Similarly flawed reasoning can be used to show that , or for example that (EMC Snapshot = NetApp Snapshot) AND (Other Vendors RAID 6 = NetApp RAID 6 (RAID DP)) .
"RAID-6 adds an additional parity block and provides approximately double the data protection over RAID-5, but at a cost of even lower write performance. As physical disks grow larger, and consequently RAID rebuild times grow longer, in some cases RAID-6 is necessary to prevent logical unit number (LUN) failure if an uncorrectable error occurs during the rebuild, or if a second disk in the array group fails during rebuild. Due to disk capacity, some vendors support RAID-6 instead of RAID-5"
NetApp RAID 6 (RAID-DP)
"RAID-DP from NetApp is a proprietary implementation of RAID double parity for data protection. RAID-DP falls within the Storage Network Industry Association definition of RAID-6. RAID-DP is also a trademark of NetApp.
Unlike traditional RAID-6, RAID-DP utilizes diagonal parity using two dedicated parity disks in the RAID group. RAID-DP is also similar to other RAID-6 implementations in terms of the reliability metrics and its ability to survive the loss of any two disks; however, a third disk failure will result in data loss. Whereas current RAID-6 implementations incur an I/O performance penalty as a result of introducing an additional parity block, RAID-DP is optimized in terms of reducing read I/Os due to the way the NetApp controller handles parity write operations. Unlike other storage controllers that write changes to the original location, the NetApp controller always writes data to new blocks, thus making random writes appear to be written sequentially. It is important to follow NetApp best practices for sizing the array to ensure a consistent level of performance for Exchange implementations"
This weekend, I was reading an interesting history about "blended" solutions (using NFS and FCP for database environments) and I'd like to talk about some points.
The author wrote: "You use the Celerra for NFS access to Oracle objects that do not need high-performance, low-latency I/O."
Probably for the NFS implementation on Celerra this is true. On NetApp you can have also a "blended" solution, but the NFS can also be used for high performance environments. Please take a look on this report:
Using 10Gb (NFS) - Average ~8Gbits/s - wire speed
Using 10Gb (NFS) - Average ~14Gbits/s - Database throughput (AWR)
Using 10Gb (NFS) - Average 5ms (latency) - Database Latency
Are the throughput and latency above enough for high performance environments using NFS?
Another interesting example about iSCSI performance. One of our customers wanted to make tests using iSCSI - Exchange environment (sorry not Oracle, but an example of database too). The customer were very concerned about the following:
Quote from the customer: "iSCSI uses 1Gbit/s and our FC infrastructure uses 2Gbits/s. We wanted to check the performance/latency using IP Storage (iSCSI) and make sure this will be the best solution for us"
Environment:
DELL 2850 – 3.60 GHz (4 processors - Intel Xeon) com 4GB RAM Using traditional Gigabit Ethernet cards (NO iSCSI HBA, NO TOE) 10,000 mailboxes Exchange 2003
NetApp FAS960C (using only one controller) Data ONTAP 7G 100G LUN (volume 200GB) - aggregate 112 disks - 72GB FC- 10K
iSCSI Software Microsoft iSCSI Software Initiator 2.03 Configuration: Using MCS (Multiple Connections per Session) Note: Using 02 (two) traditional Gigabit Ethernet cards on the server to have simultaneous I/O to "have" the same idea of 1HBA FC - 2Gbits/s (wire speed)
IOMETER - 4 workers - 64KB block size with aligned I/Os on 4KB
Throughput (MBytes/s) - 221MB/s (very close or more than one HBA FC - 2Gb/s)
I/O Response Time (ms) - 1.13ms
% CPU utilization (Server Exchange) - 18.08%
Again, as I wrote on my blog - FC vs iSCSI - An interesting history..., the discussion where NFS is for yada yada yada and FC is for bla bla bla... is a non sense discussion because:
"What is high-performance?" and "What is low-latency I/O?"
Talking about protocols, in the NetApp perspective we don't care - NetApp has FCP, iSCSI, NFS and CIFS available in the same architecture, implemented very well using WAFL.
BTW, I liked the sentence on his blog: "With EMC CLARiiON, again, a LUN is a LUN"
I believe the author forgot one very important word on his sentence...
NetApp proudly has announced the introduction of the V-Series support of Deduplication for EMC storage. Finally NetApp allows EMC customers also to have a powerful Oracle tool to simplify backups, restores and clones.
The SnapManager for Oracle (SMO) allows EMC customers (Database Administrators) to have:
1. Create fast and space-efficient backups of Oracle databases (NetApp Snapshots™). Perform instant backups with no performance hit, while applications are running. 255 backups of your database.
For example: 1TB Oracle backup in 16 seconds.
2. Reducing the mean time to recovery (MTTR) of Oracle Databases (NetApp SnapRestore®). In seconds, SnapRestore software can recover anything from an individual file, individual LUN to a multiterabyte volume so that operations can resume quickly (seconds).
For example: In a POC (Telco Company), I did a restore using SnapRestore in 23 seconds (AIX, Oracle 10G RAC, 1.4TB)
3. Saving space with clones of Oracle databases (NetApp FlexClone®). In seconds, NetApp FlexClone technology creates true clones, instantly replicated data volumes and data sets without requiring additional storage space when compared with Big Cost Volumes technologies (aka BCV).
For example: One of the our customers has been using NetApp FlexClone and they can clone their database in 26s (16 nodes, Oracle 9i RAC, 4TB)
Using NetApp technology, also Oracle Administrators can use data masking technologies. Data Masking is a solution to share production data with internal and external entities while preventing sensitive or confidential parts of the information from being disclosed to unauthorized parties.
If necessary, data can be restored from any one of the Snapshots stored on the NetApp using NetApp SnapRestore. This allows an application development team, for example, to revert to Backups (Snapshots) from various stages of their design, or test engineers to quickly and easily return data to a baseline state. Restoring to the base environment takes only seconds, and the restored environment is identical to the point at which the Backup (Snapshot) copy was created.
After the announcement, also NetApp allows EMC customers to have one technology very integrated with Oracle technology.
For example: How to "restamp" an ASM disk group name to be used on the same host? (without requiring additional storage space)
Using NetApp technology, through Oracle partnership, SnapManager for Oracle knows the ASM Disk Header format and enables ASM Diskgroup cloning via “restamping” of LUNs.
as follows. Let 
, then
) is performed, which is not an allowed algebraic operation. Similarly flawed reasoning can be used to show that 
, or for example that (EMC Snapshot = NetApp Snapshot) AND (Other Vendors RAID 6 = NetApp RAID 6 (RAID DP)) . 
![clip_image002[15]](http://blogs.netapp.com/databases/WindowsLiveWriter/clip_image002%5B15%5D_1.jpg)
![clip_image002[17]](http://blogs.netapp.com/databases/WindowsLiveWriter/clip_image002%5B17%5D_1.jpg)
