Slashdot Mirror
IBM To Purchase Informix Database
Boban Acimovic writes "According to this story on the Yahoo Financial News", IBM is going to buy Informix Database Software for $1 billion in cash. The main players in database leader struggle will be Oracle and IBM after this acquisition." That's in the commericial space - obviously SleepyCat, PostGres and MySQL and others aren't going away. And it appears that the other parts of Informix will be staying around as a seperate biz, so we should continue to see their support for OSS [?] .
For one, cache coherency on read/write conflicts between parallel cluster nodes. In 9i this will be increased to write/write conflicts as well.
other things that Oracle has that PostgreSQL may need to catchup on:
- Materialized views & snapshots
- Tons of documentation (look at the book store)
- Tablespaces and rollback segments for fine grained disk usage distribution
- 24/7 operation: the ability to take portions of the database offline for backup / recovery while keeping other parts up (i.e. tablespaces)
- Tools support (SQL Navigator, DBArtisan, etc.)
- Heterogeneous data replication
- Text-based indices (intermedia)
- XSQL and XML rowsets
And 9i is going to add even more features for 24/7 operations, such as re-creating indices without table locks, moving tables across namespaces with only short duration locks, etc.
So, while I really do like PostgreSQL, it isn't Oracle.
-Stu
IBM is no more worried about MySQL cutting into its DB2 market than Boeing is worried about Cessna cutting into its airline market.
Its not that DB2 is "Better" than Mysql any more than a 747 is "Better" than a Cessna 172, they just do different things and get used for different jobs.
Erlang Developer and podcaster
Current versions of BerkeleyDB support transactions, and note that MySQL's transaction support is built using BerkeleyDB, so clearly MySQL isn't going to support transactions and be any faster.
Hm. I do think Sybase and Microsoft are also players in the mid-to-large database market, and that a lot of companies with decent products but small market share, like Progress, would also take issue with the idea of IBM and Oracle being "it".
Sleepycat? Yeah, , Oracle and IBM do have little embedded data store products, but I'd hardly mention them in the same breath as FIlemaker, much less Oracle and DB/2. And as for MySQL and Postgres? Please. They're competition for Filemaker, MS Access, Interbase, Cloudbase and the like, and in some cases very good competition for them. But not even Postgres 7.x touches the lowest end of what the IBM, Oracle and Informix server products do. With live replication and decent hot backup features, maybe it could chew on their ankles, but that's about it. As for the middle-range, wake me up when Postgres can do clustering and failover, or when a single Postgres database can hit at least half a terabyte with good performance.
It is really dissappointing to see PostgreSQL, MySQL and SleepyCat compared to Oracle, Informix, Sybase, and DB2. The latter are enterprise databases, the former are not. While PostgreSQL adn the others are very good in the space they operate in, they do not do what Oracle and company do. To compare them as if they operated in the same space shows a gross ignorance of enterprise level data computing that is inexcusable for a site that is suppossed to be about "news for nerds." "Nerds" should know that enterprise level databases are more than transactional SQL engines (hell, in the case of MySQL and Sleepycat, not even that!).
For starters, Informix owns the telecoms market: 8 out of 10 calls placed worldwide transact across an Informix DB.
Second, big chunks of the Time Series Analysis market: several large finance houses including Merrill-Lynch, Morgan-Stanley and Chase use Informix IDS to do speeds and volumes no one else can get near. When you're doing Time Series on trillions of stock ticks per day, that's important.
Third, video: CNN, BBC, RAI, Telecinco and others use IDS and the Video DataBlade for storing video objects in the DB. CNN saves around a million per year by doing real-time ingestion and indexing of video streams, saving them on manual keying of the metadata, and getting video out onto the editors' desktops within 2 seconds.
Fourth, Data Warehousing, esp. in retail: Informix Redbrick is designed for DW, not OLTP, and it shows in the performance. Redbrick also scales to multi-terabyte far more easily than most DBs (including Informix IDS).
Informix has a sizeable, loyal customer base of people that can't get what they need from Oracle or DB2.
IBM will take the IDS/Illustra code and use it to build the next gen of DB2 with improved Object Relational support, plus star-schema support for Data Warehousing, and ride on the revenue of the installed base while they wait for the oven to go "ding". The legacy products like C-ISAM can be maintained at very little expense, giving additional long-term cash cows: it's surprising how much of that is still out there, chuggin' away untended.
Finally, there's headcount: the acquisition will also go roughly halfway toward IBM's recruitment goals for the software business, in which they intend to be one of the 3 serious players in a few years' time.
And no, I no longer work for Informix.
MasterCard International, for one. And yes, MS Outhouse is still better than Lotus Nots.
--
You answered your own question. MS isn't a player in the high-end database space. It will be.
Not to troll, or start a flamewar or anything, but MSSQL 2000 (== MSSQL 8.0 == MSSQL 7.5) is a pretty good DBMS. I haven't seen anything to touch it on a MS platform. The cynical might say that the MSSQL releases right after they hire a bunch of talent from the competition are always the best. This release appears to follow that rule.
--
Yes, the nick is flamebait