The majority of file transfer products solely support TCP/IP as their cross-platform transport mechanism. Identifying the true cost of TCP/IP is difficult and often misleading. By measuring the processor (CPU) time recorded for the file transfer task, only a fraction of the actual processor usage caused by file transfers is identified. A large portion of actual processor usage is recorded in a single bucket of time for all TCP users in the system, thus hiding the true cost of file transfers.
At best, the majority of older products use standard TCP socket programming techniques that have been available for a number of years. These techniques typically do not take advantage of several enhancements to lower CPU utilization. Among these enhancements are Communication Storage Manager (CSM) buffers and performance improvement features for Hipersockets. Additionally, some products use SNA protocols causing even more overhead.
Alebra regularly reviews new and improved methods to reduce TCP/IP overhead and provides customers with updated versions of PDM to keep processor overhead the lowest in the industry.
Despite Alebra’s industry-leading efforts to reduce TCP/IP overhead, TCP/IP may still consume significant processor resources for heavy file transfer requirements. For these environments, Alebra recommends our innovative z/OpenGate transport. In addition to blazing transfer speeds, transfers using z/OpenGate technology require 1/20th of the processor resources as TCP/IP. While actual total savings will vary depending on the size and number of file transfers, customers will save over 100 MIPS for every 100 megabytes per second of file transfer operations.
Industry terms such as Business Intelligence (BI), Data Analytics, Data Cleansing, Data Mining and Data Warehousing have different meanings to different users. Often these terms have overlapping definitions. Alebra’s uses the term Data Warehousing to mean the movement of data between an Operational Data Store (ODS) and a separate data store used for other processing. Most often, this Data Warehouse is on a different physical system likely running a different operating environment.
Over the years, several trends have emerged:
Data stores have grown exponentially is size. Databases of hundreds of gigabytes are common. Many data stores are terabytes in size (i.e. Big Data)