Click on any item to view...
Performance is Something We Like to Talk About!
A key issue for both the configuration and implementation of a host access gateway is "how does it perform?" With ScreenSurfer, the answer is simply that it is the best in the industry. The reasons for this include a variety of techical design points but are more simply described:
ScreenSurfer was designed and constructed from scratch as a server-based emulation engine with integrated scripting.
This is a key differentiator from other products that provide host access for Web applications. Many of our customers have had experiences with a variety of products that share a common trait: poor performance.
Our Competition's Red Flags
If a product is used both for PC-based GUI renovation and server-based HTML generation, it is suspect in terms of performance. Many implementations have ported PC-designed and implemented code to execute in a server environment with a variety of undesirable performance side-effects.
A classic performance bottleneck is the use of a single thread of execution in critical areas. This is necessary to support a single-PC design that has been modified to execute in a shared-logic server environment.
Another performance inhibitor found in competing products is excessive memory requirements on a per-session basis. When combined with slow code written for use on a single PC ("we don't have to worry about performance; we've got our own dedicated CPU!"), it causes serious restrictions on the number of sessions supportable on a single server.
ScreenSurfer's Green Light
ScreenSurfer, on the other hand, has key performance-contributing features which include:
- Full multi-threaded design for no bottlenecks
- Carefully managed memory allocation; minimal "dynamic heap" use
- Compiled Template Technology: HTML formatting streamed-out directly from inside the scripting engine
- Scripting integrated with protocol interpreter for low-level synchronization on screen transition events
- Low overall memory per session (below 40-100K per session depending on application data size)
The Proof is at our Customers
One customer, a large health insurance organization with a 150 person call center, recently reported that even during the heaviest use, their ScreenSurfer server rarely goes over 5% CPU utilization (using a twin-CPU 333Mhz Pentium I server). This is an environment with 1000 screens, 300 of which are enhanced through ScreenSurfer's Compiled Template Technology.
Another customer, a large university library system, has a 250 user license for web-based access that has a full redesigned web interface. This customer also rarely sees CPU utilization over 5% on a single CPU Pentium server. This is particularly rewarding since a prior (competitive) solution couldn't attain the same page-to-page performance as ScreenSurfer with far fewer concurrent sessions (and with a high cost UNIX server required!).
It's not Just the Hardware Cost
Certainly, a product like ScreenSurfer with industry-leading performance reduces hardware costs since on one low-cost server hundreds of concurrent sessions are easily supported. A more important measurement is the page-to-page performance experienced by users.
With the ability to support well over 100 transactions per second on a single-CPU Pentium server, ScreenSurfer provides the fastest page-to-page response time possible and assures that users are not slowed-down during the transition from native 3270 access to browser-based access.