Ibrahim Jaluta

Criteria for selecting software development models

The software has been playing an important role in our daily life. A need to come up with sound methodology for developing software becomes vital. There are two methodologies that are used nowadays: the heavy weight and the light weight methodologies. Both methodologies use development models such as waterfall, agile and spiral. The paper presents the two main methodologies used in software industries with their advantages and disadvantages. The paper provides criteria for selecting the appropriate model for certain applications.

Concurrent multidimensional index algorithms in page-server DBMS

In this paper, we present new concurrent R-tree algorithms for page-server database management systems. The algorithms allow for inter-transaction caching of data and index pages, so that a page can reside in a client cache even when no transaction is currently active at the client which reduces the number of client-server interactions and offload server resources (CPU and disks), thus improving client-transaction response time. Cache consistency is guaranteed by an enhanced callback-read locking protocol. Both the concurrency-control and the replica-management protocols operate at the page level.

Adaptive replica-management and concurrency-control protocols for indexes in page-server DBMS

In this paper, we present adaptive concurrency-control protocol, adaptive replica-management protocol, and B-link-tree index algorithms for page-server database management systems in which the physical database is organized as a sparse B-link-tree. In these algorithms, a transaction T at client A can update a record r in a U-locked data page (B-link tree leaf-page) P cached at A, while transactions at other clients can simultaneously cache P and fetch available records other than r in P. This would not be possible if T should X-lock P in order to update it (as in the standard concurrency-control protocols). Hence, the degree of concurrency in the page-server system will be increased.

TB±tree: Index structure for Information Retrieval Systems

Information Retrieval Systems (IR) is using different indexing techniques to retrieve information such as, Inverted files, and Signature files. However, Signature files are suitable for small IR systems due to its slow response, while inverted file have better response time but its space overhead is high. Moreover, inverted files use B±trees for single-word queries. In this paper, a new indexing structure called TB±tree to be used in the design of inverted files for large information retrieval systems. The TB±tree is a variant of the B±tree that supports single key-word queries and phrase queries efficiently. In TB±tree algorithms which represent each key-word stored in the index by a numeric value, and this numeric value can be used as encryption and inforce security. The numeric value for each keyword is stored in binary format, which may reduce the size of the index file by 19%. The records in TB±tree may be of variable length.

Enhancing concurrency and recovery in Ingres database

In Ingress spatial index (Hilbert R-link tree) algorithms, a search operation (window query or exact-match query) and an update operation (insert or delete) S-lock one page at a time during index traversal. Then the target leaf pages are locked in S mode for fetching and locked in X mode for updating. If a page split is needed, then it is propagated bottom-up using lock-coupling protocol. To adjust minimum bounding rectangles another pass over the index is needed. In this paper, we present improvements to Ingress spatial index algorithms that would enhance concurrency and recovery in Ingres database. In our algorithms, a search operation traverses the tree by latching one node at time while an update operation (insert or delete) traverses the tree using latch-coupling with U latches. Update operations and tree-structure-modifications (such as page split, merge, link, or adjusting minimum-bounding-rectangle) are executed together in one pass over the Hilbert R-link tree from the root page down to the leaf-page level. To simplify recovery, each tree-structure-modification latches two pages on a single level of the tree, executed as an atomic action, and logged using a single redo-only log record. Thus, interrupted tree-structure-modifications (TSM) is never rolled back when transaction that triggered such TSM aborts or system fails. The algorithms keep the Hilbert R-link tree balanced during normal processing and after transaction aborts (or system failure) to guarantee logarithmic search path.

Cache consistency in adaptive page-server database systems

We present cache consistency (adaptive concurrency-control and adaptive replica management) protocols for page-server database system in which the database is indexed by B±tree. These protocols avoid the data contention that may occur when concurrency control and replica management are performed at the page level. Page-level X locks are needed only for splitting or merging database (index or data)) pages, while record inserts and deletes on a data (leaf) page need only a short-duration U lock on the affected page (besides the record-level X locks). Thus, a record in a U-locked data page can be updated by one client transaction while other records can simultaneously be fetched by transactions that cache a copy of the page at other clients. Page-level S locks on data pages are not held for commit duration, but are changed to a set of commit-duration record-level S locks at the server if the page needs to be updated at another client and is therefore called back.

Highly concurrent multidimensional index for spatial database management systems

In this paper, we present highly concurrent R-link algorithms which use a novel (cheap and efficient) technique to detect node-split. A transaction S-latches one node at a time during tree traversal when fetching (window or exact-match query) or deleting spatial objects, while a transaction inserting spatial object latch-couples and U-latches two pages at a time. Insertion of new object and the triggered tree-structure-modifications (node-split, linking a child to its parent, increase-tree-height) are performed in one pass over the R-link tree. To simplify recovery, each tree-structure-modification is executed as a single atomic action involving pages on a single level of the tree, and logged using a single redo-only log record. The algorithms keep the R-link tree balanced during normal processing and after transaction aborts (or system failure) to guarantee logarithmic search path. Our approach can work for a large class of spatial index trees that are based on R-tree.

Concurrent Hilbert R-link tree

In this paper, we develop new concurrent algorithms for Hilbert R-link tree in which a search operation (window query or exact-match query) latches one node at time while an update operation (insert or delete) latches two pages at a time as in concurrent B-tree algorithms. An update operation uses latch-coupling with U latches during tree traversal. Update operations and tree-structure-modifications are executed in one pass over the Hilbert R-link tree from the root page down to the leaf-page level. To simplify recovery, each tree-structure-modification latches two pages on a single level of the tree, executed as an atomic action, and logged using a single redo-only log record. The algorithms keep the Hilbert R-link tree balanced during normal processing and after transaction aborts (or system failure) to guarantee logarithmic search path.