Archive for January 2010
As many of you know, the features and functions in the mainframe version of DB2 (DB2 for z/OS) and the distributed version (DB2 for Linux, Unix, and Windows) have been converging for a number of years.
These two flavors of DB2 now share much of the same SQL syntax. You can read the details in SQL Reference for Cross-Platform Development. There are some differences in syntax, because the platforms – hardware and software – have different strengths, and the SQL is tailored to take advantages of those strengths to deliver the best performance, scalability, availability, and security on each platform. However, the differences in syntax are being minimized, and are the exception rather than the rule.
The drivers, services, and tooling have also been converging. The Data Studio tooling provides common services for all platforms, allowing you to easily administer DB2 across all platforms.
Also, features like native XML data management (pureXML), clustering (pureScale), and compression (Deep Compression) use similar architectures and algorithms across both of these flavors of DB2. For instance, pureScale on Unix uses the same architecture and approach as the Data Sharing technology in Parallel Sysplex on the mainframe. And, of course, the Data Sharing technology is recognized as the most effective and efficient scale-out architecture in the industry. Similarly, IBM recently brought compression techniques from the mainframe to Linux, Unix, and Windows to even further stretch that flavor of DB2’s lead in storage optimization on distributed systems.
So, while these two flavors of DB2 are not identical, they are coming closer together from both a feature and a function point-of-view. Their interfaces are also converging, and currently share a high degree of commonality. This convergence is making it easier for many organizations who use both flavors of DB2 to re-use their DB2 skills on the other platform. I have had numerous conversations with DB2 for z/OS users who have easily picked up DB2 for LUW skills, and I am happy to let you know that this will become even easier as we move forward, and this convergence continues in future releases of DB2.
For another perspective on how the mainframe and distributed versions of DB2 are converging, and for my inspiration for this blog post, see the DB2 Technology Converging article in the IDUG Solutions Journal.
I mentioned earlier this week that IBM has published scale-out numbers for DB2 pureScale. I thought I’d take a few minutes to let you know about those numbers. To measure the efficiency of scaling out a system, IBM successively added DB2 servers to form larger and larger clusters. A workload that is representative of a Web commerce application (90% read and 10% write) was used. Unlike many Oracle RAC benchmarks, the IBM lab tests did not require any cluster awareness (i.e. transactions were completely random, with no need for transaction routing to specific nodes). The diagram below shows the results. Up to 64 nodes in the cluster, the scalability (compared to the 1 member result) is above 95% and at 128 nodes the scalability is at 84%.
By the way, if you want to see how easy it is to add or remove nodes in the cluster, look at the following video. Please remember that the DB2 pureScale approach does not require any re-partitioning of data, either manual or automated, when nodes are added or removed. This is because DB2 pureScale is based on the approach that has been successfully running on the mainframe for decades, which takes advantage of centralized locking to provide the ultimate database scale-out efficiency.
All Larry Ellison’s recent bluster about IBM reminded me of Don Quixote tilting at windmills. So I thought I’d try my hand at capturing that thought. As you can see, I have extremely limited artistic skills. So, apologies in advance for my rudimentary drawing abilities. But I did have some fun spending 20-30 minutes trying to put this down on paper.
Larry Ellison was his usual entertaining self on yesterday’s Webcast announcing the Oracle/Sun strategy. However, some of his statements were so far from reality that it left me wondering if he knows what’s happening in the database market. For instance, he crowed “the Oracle Database scales out, IBM DB2 for Unix does not. Let me see, how many servers can IBM put together for an OLTP application? Let’s see, how many can they group together? Um, one. They can have up to one server attacking really big jobs. When they need more capacity, they make that server bigger. And then they take the old server out, put a bigger one in. And when you’ve got the biggest server, that’s it. That’s all the can do for OLTP“. Actually Larry, you couldn’t be more wrong. DB2 pureScale supports up to 128 nodes. And, not only that, but IBM has published scale-out numbers for DB2 pureScale, something that we would love to see Oracle provide for Real Application Clusters (RAC).
He also claimed that DB2 “can’t scale out, they can’t do cloud, they can’t do clusters, the can’t do any of this“. The fact is, DB2 offers both pureScale (using a shared-disk architecture) and DPF (using a shared-nothing architecture). This allows users to choose the approach that best suits their environment. Oracle, on the other hand, offers only a shared-disk RAC architecture. Some might argue that Exadata simulates a form of shared-nothing backend for Oracle. However, this is essentially a band-aid that attempts to address scalability issues by throwing more hardware at the problem.
As regards whether DB2 can do the cloud, many people would argue that a shared-nothing architecture like DB2 DPF is the best approach for private clouds. Also, don’t forget that DB2 already has multiple cloud offerings in the marketplace, including DB2 on Amazon Elastic Compute Cloud (EC2).
Finally Larry claimed that “you would’ve thought, years ago, that IBM would have come out with a database machine. I mean its so obvious, they’ve got hardware, they’ve got DB2. Why in the world didn’t they come out with a database machine? It’s fascinating“. Actually, IBM has come out with such an offering. IBM has been doing this since we introduced the Balanced Warehouse to the market in 2007. The most recent examples of such systems that IBM has brought to market include DB2 pureScale and the IBM Smart Analytics System (which includes the entire stack needed for analytics, from hardware through ETL, warehousing, reporting, and analytics).
I really am astonished that Larry Ellison would make such fundamental factual errors on such a prominent Webcast.
This year the European IBM Information on Demand (IOD) conference is in Rome, Italy. It runs from 19 May to 21 May 2010 in the Marriot Park Hotel. To learn more about the conference, see the IBM Information on Demand EMEA Conference Web site.
IBM is currently accepting proposals for speaking sessions at the conference. IBM is particularly interested in securing DB2 users to speak. If you are a DB2 user and have something interesting to share, please submit a proposal. Interesting topics include sharing your experiences with a particular DB2 feature, best practices you observed when working with DB2, or literally anything else that you think other potential DB2 users can benefit from hearing. To submit a proposal, go to the Call for Speakers Web page.
Are you interested in deploying IBM software on a scalable, secure, and on-demand cloud environment? If so, then you will be interested in a free 2-day class that IBM is offering in the Toronto area. Attendees will receive hands-on knowledge of the AWS building blocks. You will learn about the services at feature, function, and philosophy levels that will help you to understand the intersection points between AWS and IBM. You will also learn specific details about using IBM software like DB2, Informix, and WebSphere sMash in Amazon Web Services (AWS) environments. The class covers a wide variety of technical and non-technical topics, such as:
- Cloud computing, virtualization, and AWS tools and technologies
- IBM products and cloud computing solutions available on AWS
- How to leverage existing IBM and AWS technologies to achieve Software as a Service
- Hands-on sessions with AWS technologies and various IBM products on Amazon Machine Images
- And more…
The class runs from 25 Jan 2010 to 26 Jan 2010. For more information and to sign up, see Amazon Web Services (AWS) training
If you are a DB2 user, make sure to join the recently created LinkedIn groups for the IDUG conferences. You will get all the latest news about locations, speakers, special offers, and more. Here are the links: