Emerson: DeltaV
From Devwiki
Emerson Process Management [formerly Fisher Rosemount] is a division of Emerson Electric, a $20 Billion consistently profitable engineering group. Emerson is unique in that it is the only current licensee of Objectivity/DB source code. On March 1, 2011, Emerson told us that Delta V is now a $Billion a year product. It has crossed that bar twice and slipped below it twice, but it's now on its way up again. They are considering all options for their next generation product.
Contents[hide] |
[edit] Objectivity Case History
This page needs expansion. Please update it and remove this banner when all fields are complete. |
[edit] Important
This information is an archive, so any use of the present sense in the text should be taken in the historical context, generally determinable from the Status section below.
[edit] Customer Information
- Customer: Emerson
- Project: DeltaV Historian
- Location:
- Territory:
- Industry Verticals: Process Control & Automation
- Technologies: Client-Server
- Application Domain: Process control data collection and archiving.
- Market Characterization:
- Number of developer licenses:
- Runtime license volume and type:
[edit] Status
- First Contact: 1992
- Lead came from: DEC
- Evaluation Start Date: 1994
- Evaluation Finish Date: 1994
- First Purchase Date: 1995
- Deployment Date: 1996
- Current Status: Deployed in thousands of plants worldwide.
- Can we talk about this customer and the product/project? Yes
- Referenceable?: Yes, but the project lead can only be approached via the Objectivity Sales Representative.
[edit] Environment
- Hardware: Original target was Vax, then Intel
- Operating System: NT
- Precision: 32-bit
- Development language: C++
- Compiler: Microsoft
- Third Party vendor tools: DCOM
- Open Source tools:
[edit] The Project/Product
[edit] Project Background
Rosemount Controls first approached us around 1992. At that time we were starting work on high availability features, which was their highest priority. The company merged with Fisher and it looked as if the opportunity was over. However, they started specifying the Delta V framework and conducted an extensive evaluation of Objectivity/DB, resulting in a $4M initial purchase.
[edit] Project/Product Description
Batch Historian collects process data and events from systems such as DeltaV and potentially legacy process control products. Many industries, such as pharmaceuticals, have legal requirements for maintaining this type of information for up to 7 years. Other industries use this data for quality control and process improvement operations. The core of Batch Historian is a C++ application that records the results of a batch process control run. The second main piece of Batch Historian is the reporting side, built around Objectivity’s SQL++ server, containing a variety of stored procedures linked in. The end reporting clients use ODBC connections to the SQL++ server to generate the required reports.
SQL++ and stored procedures provide an effective means of speeding up reports by using a time-based index lookup to dramatically narrow the volume of data that must be processed on most queries.
[edit] Buying Criteria
[edit] Business Priorities
- Company stability
- License prices - Fisher placed an initial $4M order.
[edit] Technical Priorities
- Reliability
- Scalability: Batch Historian is required to save the information from a batch process control run for many years. The resulting volume of data can grow quite large.
[edit] Difficulties Encountered
[Caveat: This was written around 1996. The problems were solved and Fisher became a very good reference for us, though current status should be checked with the Objectivity Rep.] Fisher has stressed our SQL interface more than any other I am aware of, and as a result, have uncovered several problems in it. In general they perceive Objectivity’s SQL interface as a less mature, less important to Objectivity language binding that gets less attention than our other language bindings, and is typically even released later. 5.0 was the first release that the intent was to release SQL in concert with C++, which was appreciated, but even there, the maturity was significantly behind the C++ side and felt like “beta” to them.
Some of the more significant problems encountered:
- Memory leaks
- 14 character limitation on ODBC name (months to debug)
- Memory initialization errors when using stored procedures- required workarounds such as adding a dummy column
- Single-user concurrency, i.e. one read-only query would block another read-only query
- Result sets greater than 5000 rows would crash the system
- Security control did not work- i.e. all clients had to have same access privileges to work
March 2011 Update:
- SQL++ join table operations, sometimes amounting to millions in a day. They are still using Objectiviy/SQL++ Release 6, which does not take advantage of names associations.
- Lock server bottlenecks. The newer lock server has multithreading. They could also move to partitions and multiple lock servers.
- Rolling upgrades will be necessary in the future.
- Upgrading the schemas and catalogs for existing sites will be interesting.
[edit] Competitors/Alternatives
- ObjectStore
- Home grown solution.
[edit] Why They Chose Objectivity
- Objectivity/DB was originally used in a prototype framework for the Delta V project, which is the prime monitoring target for Batch Historian.
[edit] Partners
- Digital Equipment Corporation (DEC), which sold Objectivity/DB as DEC ObjectDB at the time.
[edit] Collateral
- Press Releases:
- Fliers:
- White Papers:
- Case Study:
- Other:
