Tuesday Nov 11, 2014

Software in Silicon, Innovations for Database

Dear Partner,

Oracle Open World 2014 still give us updates and news to share with you.

Please have a look to this video, where Juan Loaiza, Sr. VP of Technology at Oracle, gave an overview of innovative new Software in Silicon technology and its benefits to databases and applications

Thursday Oct 09, 2014

Oracle Launches the Oracle FS1 Series Flash Storage System

Dear Partner, 

 following Oracle Open Word announcements, we are proud to spread the news that Oracle has launched the Oracle FS1 Series Flash Storage System, innovating with new QoS Plus software and provisioning one-click application storage.

Please have a look and read the entire article HERE, you can read how the new Oracle FS1 Series flash storage system scales to petabytes of all-flash capacity while delivering a multitude of innovations designed to take flash technology to new extremes.

Sunday Aug 03, 2014

Good News around ZS3

"Sometimes we all miss changes that are happening right in front of us. Aspects of Oracle's recent ZS3-2 announcment made me realize this...." Mark Peters about ZS3

Dear Partner,

Following the recent ZS3 announcements, the ESG analyst Mark Peters posted a blog post in which he explains the perceptions he held about Oracle storage and how the ZS3 is making him reconsider those - highlighting the fact that with the ZS3, Oracle is proving itself as an open, high performance at low cost ("we're not talking silly money to get the job done well"), real convergence storage player that is delivering on the ZS Series with "prowess and investment".

You can view the full blog entry here: http://www.esg-global.com/blogs/oraclee28099s-new-zs3-2-makes-me-reconsider/

Thursday Jul 31, 2014

What is the ideal Data Protection platform? Webcast replay available!

On Thu. 24th of July Oracle hosted a webcast explaining how Oracle’s Application Engineered Storage is the ideal Data Protection platform for Oracle’s Engineered Systems. The recording of this webcast is now available here.
The webcast slides are also available on the community workspace storage folder.

Remember, you need to be a registered member of the community to be able to access the workspace.
If you are not a member yet, you can register clicking here.

Find out how the unique integration features between Oracle’s Engineered Systems & Oracle’s Application Engineered Storage can help you help your customers, discover new opportunities and stay ahead of the competition.

We look forward to welcoming you to another webcast soon!

Wednesday Jul 30, 2014

Oracle ZFS Storage ZS3-2 Sets World Record on SPC-2/E Benchmark

Dear Partner,

Recently Oracle published a new SPC-2/E world record benchmark for the ZS3-2 storage platform demonstrating, once again, that the Oracle ZFS Storage ZS3-2 offers customers excellent price/performance for high-throughput workloads such as data warehouse, business analytics and video on-demand applications.
The ZFS Storage ZS3-2 beats IBM DS8870 by 10x(2) and HP P9500 by 7x(3) on price/performance.

For further information, please refer to the press release regarding this benchmark: http://www.oracle.com/us/corporate/press/2226303

Friday Jun 20, 2014

Webcast Replay: Why DBAs care about Storage?

Dear partners,

On Wed. 18th of June we hosted a webcast highlighting how Oracle’s Application Engineered Storage can make the everyday life of a DBA easier by automating or simplifying many of the tasks associated with managing a database (troubleshooting, tuning, backups, etc.). The recording of this webcast is now available at: http://w.on24.com/r.htm?e=806789&s=1&k=5581770600654C7696766CFF7569E7B4

Find out how the unique integration features between the Oracle database & storage platforms can help you help your customers, discover new opportunities and stay ahead of the competition.

We look forward to welcoming you to another webcast soon!

Tuesday Jun 10, 2014

Oracle ZS3 Contest for Partners: Share an unforgettable experience at the Teatro Alla Scala in Milan

Dear valued Partner,

We are pleased to launch a partner contest exclusive to our partners dedicated to promoting and selling Oracle Systems!

You are essential to the success of Oracle and we want to recognize your contribution and effort in driving Oracle Storage to the market. To show our appreciation we are delighted to announce a contest, giving the winners the opportunity to attend a roundtable chaired by Senior Oracle Executives and spend an unforgettable evening at the magnificent Teatro Alla Scala in Milan, followed by a stay at the Grand Hotel et de Milan, courtesy of Oracle.

Recognition will be given to 12 partner companies (10 VARs & 2 VADs) who will be recognized for their ZFS storage booking achievement in the broad market between June 1st and July 18th 2014.

Criteria of Eligibility

  • A minimum deal value of $30k is required for qualification
  • Partners who are wholly or partially owned by a public sector organization are not eligible for participation 


  • The winning VARs will be:
    • The highest ZS3 or ZBA bookings achievers by COB on July 18th, 2014 in each Oracle EMEA region (1)
    • The highest Oracle on Oracle (2) ZS3 or ZBA bookings achievers by COB on July 18th, 2014 in each Oracle EMEA region
  • The winning VADs (3) will be:
    • The highest ZS3 or ZBA bookings achiever by COB on July 18th 2014 in EMEA
    • The highest Oracle on Oracle (2) ZS3 or ZBA bookings achiever by COB on July 18th 2014 in EMEA 

The Prize

Winners will be invited to participate to a roundtable chaired by Oracle on Monday September 8th 2014 in Milan and to be guests of Oracle in the evening of September 8th, 2014 at the Teatro Alla Scala. The evening will comprise of a private tour of the Scala museum, cocktail reception at the elegant museum rooms and attending the performance by the renowned Soprano, Maria Agresta. Our guests will then retire for the evening to the Grand Hotel et de Milan, courtesy of Oracle.

Oracle shall be the final arbiter in selecting the winners and all winners will be notified via their Oracle account manager.
Full details about the contest, expenses covered by Oracle and timetable of events can be found on the Oracle EMEA Hardware (Servers & Storage) Partner Community workspace (
FY15 Q1 ZFS Partner Contest).
: access to the community workspace requires membership. If you are not a member please register 

Good Luck!!

For more information, please contact Sasan Moaveni.

(1) Two VAR winners for each EMEA region – Eastern Europe & CIS, Middle East & Africa, South Europe, North Europe, UK/Ireland & Israel - as per the criteria outlined above
(2) Oracle on Oracle, in this instance, means ZS3 or ZBA storage attached to DB or DB options, Engineered Systems or Sparc servers sold to the same customer by the same partner within the contest timelines.
(3) Two VAD winners, one for each of the criteria outlined above, will be selected from across EMEA.

Thursday Apr 17, 2014

Speed, Simplicity, Saving. ZFS Storage Appliance

Your business can run faster and your IT operations can be more streamlined with the right storage.
Get the facts about Oracle’s latest generation ZFS Storage Appliance, the ZS3 Series.

See the infographic video here: http://medianetwork.oracle.com/video/player/3472158176001

Get better performances with Oracle ZFS Storage Appliance: ask us for further informations!

Monday Feb 10, 2014

Database Speed & Efficiency: the webcast

Dave Vellante (Wikibon Chief Research Officer) and Steve Zivanic held a webcast On Feb. 6th, highlighting the business benefits of ZS3 with Oracle Database, followed by an industry overview and competitive comparison of ZS3 vs. NetApp FAS6000 systems.

The public webcast replay is available at this link.
The presentation materials used by Oracle and Wikibon during the webcast are available here.

Ask us how to join to the community!

Wednesday Feb 05, 2014

Making The Right Hybrid Storage Array Buying Decision: Oracle Zs3 Series Stands Out As A Hybrid Enterprise Storage Solution

In a new Head-to-Head Product Report on Hybrid Storage Arrays, the Data Center Infrastructure Group (DCIG) concluded that "the EMC VNX5400 is probably the most ill-equipped of these arrays to meet enterprise performance demands short and long term".
The report goes on to say that "the Oracle ZS3 Series stands out in stark contrast to other offerings from enterprise storage providers. It gives enterprise organizations the features they need in a hybrid storage array without having to compromise."

What to say more? Just review the full report HERE

Wednesday Jan 29, 2014

Meeting Engineered Systems Backup Needs: "Best Practices for Exadata Backup Using Oracle ZFS Storage" Webcast incoming

Next 4th February will be hosted a technical webinar, designed for DBAs and other IT staff using or considering Oracle Engineered Systems.
The two speakers  (Travis Parker -Sr. Database Administrator, The Nielsen Company- and Nancy Hart -Director, Oracle-on-Oracle & Application Integration for Storage, Oracle- ) will discuss around best practices for reducing backup windows, better meeting stringent recovery objectives, and more.

Join us and learn how to quickly implement and tune your backup environment, avoid the risk and costs associated with third-party backup and deduplication systems and maximize the return on your backups with other secondary processes on Tuesday 4th February at 8am PT/ 11am ET.

Find all the details of the webcast here.

Tuesday Jan 28, 2014

Run Oracle Database Like a Ferrari, But Save Like a Prius with Oracle Storage

It's a well-known fact that only Oracle storage is co-engineered with Oracle Database to deliver unique capabilities that maximize performance, efficiency and ROI. Third-party storage systems simply can’t match.

Next 6th of February we'll hold a webcast in collaboration with Dave Vellante (Co-Founder and Chief Research Officer at Wikibon). Steve Zivanic (Storage Business Group VP, Oracle) will be a speaker too: listen to them and you will know the advantages of running Oracle Storage Systems on Oracle Database and the tangible business benefits that brings to customers.

The webcast is particularly dedicated to those that want to learn about Oracle Database on Oracle Storage and how the combination delivers significant cost savings, performance and capabilities unavailable elsewhere.

To learn more, we invite you to join our webcast on Thursday, 06th February 2014 at 11 am PT/ 2pm ET.

You can find all the details of the webcast here.

Thursday Jan 23, 2014

StorageTek T10000D Wins Industry Awards

We're proud to announce that Oracle's StorageTek T10000D tape drive has been declared the WINNER in the "Visionary Professional Media & Entertainment Class Archive Storage" category at Storage Visions 2014

These awards recognize companies advancing the state of the art in storage technologies utilized in consumer electronics, the media and entertainment industries, visionary products for the digital content value chain and digital storage media and entertainment end users.

But it's not finished yet!
StorageTek T10000D tape drive is finalist as Storage Magazine/SearchStorage.com "Product of the Year 2013 Award" for backup hardware. Let's wait for the results in February 2014.

More details available at the Oracle EMEA Partner Community for HW - click here!

Tuesday Sep 03, 2013

Optimized Storage for DB Workloads (part 2)

In a previous issue of this blog, we have analyzed why the Oracle Sun ZFS Storage Appliance is an excellent storage system to host an enterprise database. Because of its unique controller architecture based on the Hybrid Storage Pool (HSP), it guarantees best in class performance and ease of integration. In this second part of the blog, we are going to address another critical requirement in today’s Enterprise database implementations: compression.

Best in class compression capabilities is a must if we consider the current growth of structured data in the IT centers. While traditional compression algorithms typically found in disk subsystem controllers can do a good job, a more specialized compression mechanism aligned with the database structure can go much further in terms of data reduction.

At the top of that, compression is not just a matter of saving space in a database environment:

  • It also has a great side effect when properly implemented: it reduces IO traffic between storage and the database node. The point is that disk drive capacity and IO capability are quite unbalanced. We can store massive amounts of data in today’s HDD, but the IO capabilities of one given disk have remained basically unchanged in the recent past. At the same time, the more data we store, the more IOs we need from a given disk. Therefore, it is often more important to pay attention at decreasing the IOs from a storage system rather than increasing its capacity.
  • It increases efficiency not only in terms of data reduction on disk, but also for the backup tasks, Disaster Recovery replication, memory usage (SGA) and ARC/L2ARC caches when using the Sun ZFS Storage Appliance. On that aspect, it goes far beyond what a traditional compression or deduplication approach at disk controller level can achieve.

Compression has always been an important aspect of an Oracle database environment. The table below summarizes the different types of compression which are available in an Oracle database and provides for each of them:

  • the database level needed to use the compression capability;
  • what is required in terms of storage or database option to use it;
  • what type of compression it operates;
  • what type of operation the compression applies to.

Compression DB Level   Requirements Type of Compression  DB Operations 
Index Compression 8i 

  • All types of storage
  • Part of the core product, no specific option required 

Row compression on indexes through deduplication of the index leading columns Compression on indexes when created or rebuild an index to compress it.
Basic Table Compression 9i
  • All types of storage
  • Part of the core product, no specific option required
Row compression on tables through deduplication of repeated values Only applies to bulk operations, i.e. compression is not maintained during DML operations such as INSERT, UPDATE and DELETE, so basic compression should be reserved for data which is mostly read-only.
Advanced Row Compression** 11gR1
  • All types of storage
  • Requires the separately licensed ACO*
Row compression at block-level through deduplication of repeated values Compression is maintained during DML operations (INSERT and UPDATE) with some restrictions (compression is triggered when new records reach a threshold).
RMAN Backup Compression 10g
  • All types of storage
  • Requires the separately licensed ACO*
Compression levels are BASIC (BZIP2), LOW (LZO), MEDIUM (ZLIB) and HIGH (BZIP2) Backup data is compressed before being written to disk or tape and read back directly in compressed format (only for backup sets).
SecureFile Compression and Deduplication 11gR1
  • All types of storage
  • Requires the separately licensed ACO*
Compression is based on industry standard compression algorithms Deduplication is performed first on SecureFile data, then compression and finally encryption if required.
Data Pump Export Compression 11gR1
  • All types of storage
  • Requires the separately licensed ACO*
Data Pump compression provides comparable reduction to standard gzip algorithms Inline compression during export and import of dump files.
Data Guard Redo Transport Compression 11gR1
  • All types of storage
  • Requires the separately licensed ACO*
Data Guard Redo Transport Compression uses the zlib compression engine at level 1 (equivalent to gzip -1 level) Data Guard Redo data is transmitted in a compressed format to reduce network bandwidth and improve performance.
Hybrid Columnar Compression 11gR2
  • Only available on Exadata, ZFS Storage Appliance and Pillar
  • No license required
Hybrid row/column compression that comes in 4 flavors: Query Low (LZO) and Query High (ZLIB) for warehouse compression, Archive Low (ZLIB) and Archive High (BZIP2) for archive compression Hybrid Columnar Compression is used on bulk operations. When DML is performed on HCC compressed data, new or modified rows are stored uncompressed or possibly compressed by OLTP Compression if the option is available. HCC is intended for dormant or possibly near-dormant data.

Advanced Compression Option (ACO) in Oracle Database Enterprise Edition. We do not mention in the table the following ACO options: Flashback Data Archive (11g), Heat Map (12c), Automatic Data Optimization (12c), Advanced Network Compression (12c), Storage Snapshot Optimization (12c) and Online Move Partition (12c).

** Starting with Oracle Database 12c, OLTP Table Compression is called Advanced Row Compression.

The Advanced Compression Option in the recently announced Oracle 12c brings several new compression features including Heat Map and Automatic Data Optimization (ADO). Heat Map is a new Oracle Database feature that collects usage information at the block and segment levels, enabling further automated compression and movement of data which becomes less accessed and used over time. Automatic Data Optimization is the policy engine that performs compression and storage tiering of table partitions according to the data maintained in the Heat Map. Compression can be any type of compression supported by the Oracle Database including Advanced Row Compression (formerly called OLTP Table Compression) and all levels of Hybrid Columnar Compression (HCC).

The Advanced Row Compression introduced in Oracle DB 9iR2 was a great improvement compared to the previous Basic Table Compression as it supports conventional DML (Data Manipulation Language) operations such as INSERT and UPDATE. Furthermore, the database does not require data to be uncompressed and keeps data compressed in memory. This means that performance can be improved on reads since fewer IOs are performed while memory can be more efficiently used. It must be noted that Advanced Compression Options (ACO) regroups several different types of compression: Advanced Row Compression, RMAN Backup Compression, SecureFile Compression, Data Pump Export Compression and Data Guard Redo Transport Compression (as well as several new options introduced in Oracle Database 12c).

Hybrid Columnar Compression (HCC) goes to the next level in terms of compression ratios compared to the other compression techniques. Typically, you will get a compression ratio between 1:5 and 1:20 as opposed to 1:2 and 1:4 for Basic Table Compression and Advanced Row Compression. HCC can be used with two compression modes: Query and Archive. The Query compression mode will provide faster decompression for data which is relatively frequently accessed but a lower compression ratio. On the other hand, the Archive compression mode will provide slower decompression for data which is not often accessed but a very significantly higher compression ratio.

Let’s now have a deeper look at how Hybrid Columnar Compression works and why it can bring so much value for compressing warehouse and archive data on Oracle’s storage.

Imagine that we have a 7-row and 6-column table as shown on the left and that we want to compress it with HCC. The first thing to understand is how such a table would be stored in an Oracle database without any compression applied. The Oracle database splits table data into “blocks”, each block containing data in the row format, i.e. the first row is sequentially written into the block, then the second one is appended and so on. If all the rows do not fit into one block, no problem, a second block is simply chained to the first one. So our table is going to be stored in the Oracle database as shown below (we ignore here the details about the header blocks and other subtle implementation tricks):


Such an arrangement of data in the database blocks is in most cases very effective when one wants to read for instance a complete row since the column values can be sequentially read for that particular row. However, it does not land itself well to traditional industry standard compression algorithms such as LZO and ZLIB since data with similar types and values are most of the time sorted in columns and are now dispersed within the blocks as show below:


This is where HCC is coming into play… In order to perform an effective compression without losing the benefit of having the rows “locally stored” within the blocks, we are going to store the data in a columnar format, but only for a subset of rows and not the whole table. So let’s say that, for the sake of our example, we define a set of 3 rows, that we call a Compression Unit (as shown on the right). For that Compression Unit, we reorganize the data in the “columnar” way, i.e. we group first the data by columns then we move along the rows. Here is what our compression unit looks like after the data reorganization process:



In reality, a Compression Unit would include thousands or tens of thousands of lines, but for the sake of our example, we are going to limit ourselves to three… Now, before storing the Compression Unit data into the database blocks, we are going to compress it. Guess what, we are now in a much better position to achieve a good compression ratio since similar data from the same columns now sit next to each other, at least for the rows that constitute a Compression Unit. After compression using a traditional “zip-like” algorithm, here is what we are going to get:


Our Compression Unit has dramatically shrunk and is now ready to be stored into the database blocks as shown below. If the compression unit does not fit in one block, it is simply split on a second block and the blocks are chained together.


In order to understand how DELETE, UPDATE and INSERT commands are managed with HCC compressed format, let’s first see in more details how a compression structure is stored within blocks:


The Header stores the offsets and lengths of all the column entries within the Compression Unit while the Bitmap is in charge of identifying the table rows which are stored by the Compression Unit (one can see the Bitmap as a “map” that gives the status of each of the row and column of the table zone which is covered by the Compression Unit). So here is how the data is actually stored into the blocks (we also represent here the Block Headers but not the Row Directories nor the Row Headers):


Let’s say now that a table row is actually deleted through a DELETE DML statement. Since the data corresponding to that row is deeply mingled with the other rows into the Compression Unit, there is no way we can free that space into the blocks. So all what we do is to mark the row data as deleted into the Compression Unit Bitmap as shown below with the small red squares into the Bitmap:


When compressed with HCC, a deleted entry is actually never removed from the table stored on the disk. It is simply marked as invalid. The only way to get rid of the data which is not more valid is to copy the table through a database command.

If a row is updated or inserted respectively through an UPDATE or INSERT statement, then the new row is simply added as a row (and not as an HCC hybrid structure) into the free space of the last block that constitutes the Compression Unit. No compression is applied to this modified or new row, except if the Advanced Row Compression option is set for the particular table or partition the row belongs to (but this is not an HCC compression which is used here). The figure below depicts the case of a row update where the old data is marked as invalid into the Compression Unit bitmap and where the new row is added in an uncompressed format into the free space of the second block (if there is not enough free space into that block, another block is simply chained to it):


Now that we understand how DML operations are performed in an HCC compressed table or partition, it becomes clear that we should not perform too many changes on that type of table and that HCC must only be used for dormant or near-dormant data such as Data Warehouse and Database Archive environments. Despite this limitation, in a vast majority of cases, the compression ratios achieved by HCC will be much better than what can be achieved through Advanced Row Compression.

Let’s now have a look at the different HCC options that can be used:

HCC Option Target Environment Typical Compression Ratio Objective
QUERY LOW Warehouse Compression Between 1:5 and 1:10 Optimized for speed where load time is important
QUERY HIGH Warehouse Compression Between 1:10 and 1:15 Optimized for speed on data access
ARCHIVE LOW Online Archive Compression Between 1:10 and 1:20 Optimized for highly compressed historical and archive data
ARCHIVE HIGH Online Archive Compression Between 1:15 and 1:50 Optimized for maximum storage reduction for data which is lightly accessed

At the top of storage reduction, HCC often brings an additional value in terms of performance for database loads which can be limited by an IO bottleneck between the database server and the storage device. In the case of database tables stored in a ZFS Storage Appliance and HCC compressed, all the queries to the disk subsystem are returned in the HCC compressed format, resulting in a much faster transmission of the data and greatly improved query performances. This is especially true when the ZFS Storage Appliance unit is connected to the database node through a low latency Infiniband network, which represents quite a unique capability of the ZFS Storage Appliance on the database storage market.

The figure below summarizes the benefits of handling HCC compressed tables:

  1. Data remains compressed during transfers over the network;
  2. Data remains compressed in the database SGA Buffer Cache;
  3. Data remains compressed in the Smart Flash Cache of the database server;
  4. Data stays compressed when handled by RMAN for backup to disk or tape;
  5. Data remains compressed when shipping logs to the Data Guard node over the network.


It is very clear that using a traditional disk array based compression does not bring any of the advantages mentioned above, and this makes HCC an extremely powerful feature for storing efficiently Warehouse and Archive tables or partitions. When combined with the unique capabilities of the Oracle Sun ZFS Storage Appliance (IO bandwidth, throughput, snapshot and cloning, connectivity and observability), Hybrid Columnar Compression represents an unmatched tool to reduce infrastructure costs and improve performance.

Monday Apr 15, 2013

New analyst report on ZFS Storage Appliance

Industry analysts from Storage Strategies Now recently released a report comparing Oracle's Sun ZFS Storage Appliance to NetApp, EMC VNX and Isilon.  

The report is available now at:  http://www.oracle.com/us/products/servers-storage/storage/nas/oracle-storage-update-ssg-now-1933211.pdf.