Download File
Maybe it’s the end of the year that brings retrospective thought, or perhaps it’s one too many post-holiday feast chats over eggnog, but in any case, it is the habit of many an analyst and journalist to think through all the events of the previous year and guess how the …
While much has been written about XML data storage, none of the research to date has focused on the key questions:
In attempting to understand the answers to the above questions, ZapThink realized that there was no single answer for any of these key questions. In fact, it seems that the various requirements for XML data storage pull end-user customers in different directions when they are deciding how to implement XML storage requirements. As such, this study seeks to do what no other study before it has done: show that XML data storage is not a distinct market segment, but instead a functionality requirement for applications that require XML storage in order to achieve their overall system objectives.
“Instead of focusing on a general-purpose store for XML data, as these companies have done in the past, many XML storage vendors are moving to more purpose-built data stores that are focused on solving specific problems where XML data storage can add value,” said Ron Schmelzer, an analyst with ZapThink LLC, in Cambridge, Mass. “In the case of Excelon, they are moving toward the management of XML documents that are used in [business-to-business] interactions.”
Read more at: eWeekXML is pervasive. In a matter of years, it will fuel every application, device, and document found in enterprise networks. However, as XML proliferates, it will stress existing systems and enterprise budgets to their breaking points. This is because existing n-tier software architectures and legacy infrastructures were not designed to process this verbose new data type efficiently. What enterprises need is a new way to process XML in the network, rather than in software at the database, application server, or presentation tiers. Yet today’s existing network infrastructure is limited to switching lower layer protocols and is unable to detect XML – much less parse and process it.
An emerging class of hardware-based XML-aware network devices addresses the performance, security and management issues that come with XML’s use in enterprise applications. These purpose-built network devices enable enterprises to process high volumes of XML in a way that offers high performance, optimal security, lowest total cost of ownership, and greatest centralization of XML processing capabilities.
Ipedo has developed a proprietary, hierarchical, in-memory database engine aimed at storing XML documents "natively". The Ipedo XML Database uses memory techniques to get substantial performance gains versus some other NXD approaches. Built as an all-Java server, the Ipedo system is meant to be easily integrated with typical application server environments. Ipedo has extended XPath for search across multiple XML documents and has also included XSLT transformations within the data store itself.
The only true solutions for enterprise-wide XML data storage and retrieval are XML-enabled RDBMS and Native XML Data Store (NXD) approaches. The challenge with XML-enabled RDBMS is that they are not very well suited to highly extensible and flexible documents, while the challenge with NXDs is that they are unproven technologies with very small user bases. However, there is another option that is a sort of "middle ground": the use of an interpretive middle layer over standard RDBMS systems that offers schema-independent, "native" XML data storage. In this vein, B-Bop offers their Xfinity Server as a way of providing the features of Native XML storage while utilizing existing relational storage architectures.
The most valuable attribute of Native XML Data Stores (NXDs) is their ability to store arbitrary and highly variant XML documents. XML-enabled RDBMS systems require explicit mappings to XML documents, and by their very nature are unable to deal with XML documents that have a highly variable structure and take advantage of XML’s extensibility capabilities. X-Hive provides a solution to this problem by presenting a highly scalable NXD system that is capable of supporting a large quantity and volume of XML documents. X-Hive/DB is differentiated itself by its focus on high volume XML data storage requirements, support for advanced XML query and storage specifications, and focused support of its European customer base.
Rather than being forced to map XML documents to an alternate data representation format or insert an entire XML document into a single field as is the case with XML-enabled RDBMS, the Native XML Data Store (NXD) can accept XML documents for storage and retrieval without any modifications, mapping, or transformations. NeoCore provides a patented NXD approach that assigns a unique fingerprint called an "icon" to each XML document and then applies hashing and storage techniques to these iconic representations. This approach allows the user to experience speed and indexing improvements over other methods of XML storage and retrieval and exploits XML structure by self-constructing the data store, reducing development time.
Banks, brokerage houses, and insurance and real estate companies spent $500 million on XML-related technologies last year and will invest an estimated $8.4 billion a year by 2005, more than any other industry, according to a recent XML market research report, “XML in Financial Services,” by Ron Schmelzer, an analyst at ZapThink LLC, a market research company in Waltham, Mass.
Read more at: eWeek
SOA Implementation Roadmap