Usage Note 45037: SAS® Drug Development 3.5 - Network performance considerations
 |  |  |  |  |
Introduction
SAS Drug Development is a web-based platform that can be hosted, deployed, and configured in different ways. As a result, many outside factors can impact SAS Drug Development’s performance for end-users. This paper outlines a few of the major contributors and suggests basic tests for evaluating performance for a site’s end-users. It is important to understand these factors and their effect on SAS Drug Development’s Internet performance for end-users early in the implementation process in order to configure the solution to achieve its highest potential for the company.
A few factors that can influence performance are directly related to SAS Drug Development, but most are related to a customer’s network, Internet access, and the distance between the hosting machine and the user. This paper focuses on identifying factors that can affect transmissions in both its route and speed between the hosting and end-user machines. Because each company’s network is unique, this paper should be used as a starting point for, more technical conversations geared to the specifics of the customer’s network. Performance Factors There is a significant difference between viewing a common web page and downloading (or uploading) large amounts of data within SAS Drug Development. The factors that affect the performance of a software program like SAS Drug Development for end-users are interrelated and can therefore complicate determining the root cause of degraded performance. They include:
- network bandwidth
- latency or delay
- Internet load (number of users and Internet saturation)
- network configuration
- PC requirements/configuration
- size of data
Network Bandwidth
Network bandwidth (or data-transfer rate) consists of two major components: Local Area Networks (LANs) and Wide Area Networks (WANs). While a LAN is typically a single site’s network, WAN connectivity represents the connection to remote offices and/or the Internet. The bandwidth for both LANs and WANs is typically measured in megabits per second (Mbps) or megabytes per second (MBps).The majority of modern LAN bandwidths are either 100 Mbps or 1,000 Mbps (1 Gigabit per second). Due to the provider cost, WAN connectivity is usually much slower than the LAN. The most significant impact on performance of SAS Drug Development is the amount of bandwidth available to the Internet. The SAS Campus in Cary, NC, has around 400 Mbps of total bandwidth to the Internet. By comparison, a typical T1 line accommodates only 1.5 Mbps of bandwidth.
Without knowing the specifics of actual daily usage, it is difficult to state how much available bandwidth is needed for SAS Drug Development to deliver acceptable performance. In this case, available bandwidth is defined as your company’s Internet speed, minus your company’s used speed. The difference should be at least 1.5Mbps per SAS Drug Development user from a given corporate location.
SAS Drug Development can run at speeds less than 1.5Mbps, but performance might suffer. A good understanding of a site’s available bandwidth and load already in place on this bandwidth prior to implementation is vital when considering this variable.
-
Common Connection Types Maximum Connection Speed
- 56K Dial Up Modem - 56 kbps
- Standard T1 Line - 1.5Mbps
- T3 Line - 45Mbps
Latency
Latency, or delay, is a measure of the time it takes for a single packet of information to cross a network path one way from sender to receiver. Latency is usually measured in milliseconds (ms). Many factors can affect latency across the Internet. The most significant contributor is the physical distance between two locations: the longer the distance, the greater the latency. A quick way to demonstrate this is to open a command prompt and type
-
“ping www.google.com”.
Internet Load
The load on a location’s Internet connection has a major impact on SAS Drug Development’s performance. Load in this situation is defined as the amount of available bandwidth. The number of users and the resource intensity these users place on the connection are always significant factors when evaluating performance. With an Internet-based solution, it is not only the number of users using the solution, but more importantly, the number of total users sharing the same Internet connection. A site might have only 10 users accessing SAS Drug Development, but 500 total users sharing the same Internet connection. Network oversaturation causes TCP window delays and retries. Basically, the performance reduces in a non-linear scale (closer to exponential) and begins to degrade before 100% utilization has been reached. To illustrate this point, imagine cars on a four-lane freeway entering a section in which three of the lanes are closed for construction. Because there are fewer lanes, everybody must merge and slow down, eventually reaching speeds well below what a normal single lane would support.
The types of tasks users are performing over the Internet also have an impact on SAS Drug Development’ performance. Accessing basic HTML web pages is not a very resource-intensive task. In contrast, accessing a webcast or other similar streaming website uses additional resources and consumes considerably more bandwidth, competing with access to SAS Drug Development and resulting in reduced performance.
Network Configuration
A site’s internal network configuration can also affect performance. This paper does not discuss this topic in much detail due to the extreme variances between and complexities within company networks. Instead, a very high-level view of its effect on SAS Drug Development is surfaced in this document.
A core issue in Internet performance is the number of router hops network traffic has to travel through the Internet. The more router hops, the slower performance can be. If a network is connecting to a central office before accessing the Internet, the multiple routers between the start and end points, along with the shared Internet bandwidth at the central office might adversely impact performance compared to a similar direct Internet connection.
In addition, the site’s proxy server and firewall can further degrade performance. A firewall processes each packet sent through it and compares the packet to the access rules that are defined. If the HTTPS protocol is not allowed out through the firewall, the application does not work at all. If the firewall is configured to analyze the packets in detail, or slow in comparing requests to defined rules, added delay occurs for each packet leaving the site.
SAS Drug Development is a client (JRE)/server-based application. Most actions performed within the application initiate multiple server calls. Due to the nature of the communication; specifically the encryption and dynamic content of the interchanges, the caching, and filtering capabilities of the proxy server are not possible. However, the additional routing of the traffic and burden on the proxy server still affect the client/server conversation and degrade end-user performance.
Finally, the Internet Service Provider (ISP) can affect available bandwidth and network latency. A primary service provider with entry points at multiple Network Access Points (NAPs) on the Internet delivers higher performance than a third or fourth-party service reseller. The ultimate goal of the connection between the end-user and the hosting server is to minimize the number of hops and deviations in the network path.
PC Configuration
Because most companies use relatively modern computers, SAS Drug Development’s consumption of personal computer (PC) resources is generally not a concern. However, SAS Drug Development does use some of the PC resources, including CPU. Internet Explorer and the Java applets also consume PC memory. Typically, most end-users have sufficient processors and ample memory so that resources are not an issue. However, computers more than 4-5 years old could negatively impact access to SAS Drug Development.
For optimum performance, there are several parameters that should be set on each PC that accesses SAS Drug Development. These include proxy setups and page refresh rates and can be discussed further with your SAS Drug Development account administrator.
Size of Data
As with most other programs, the amount of data moved and manipulated have an effect on performance. Performance differs when uploading a 100 MB data set and a 50 KB data set. The rate of uploading data is subject to the previously discussed factors. Internet Performance Testing SAS provides an Internet accessible URL to test the average upload and download speeds of a network and measure latency. The results are based on the total state of all networks involved at the moment the test is performed.
SAS asks customers to perform multiple speed tests per day (mid-morning, noon, and mid-afternoon) to provide as variety of results based on normal workday Internet loads. In addition to the 3 daily tests, SAS requests the customer perform this testing three times during one week on different days to create a robust set of performance metrics.
Testing SAS Drug Development Performance through a Customer’s Internet Connection
After a general review of a site’s Internet performance is conducted, a performance test using SAS Drug Development from the customer’s site is implemented.
Once the Internet and SAS Drug Development performance results are collected, SAS can compare the customer site Internet performance results against the performance from a SAS regional office in the same area. Comparisons between these two sets of metrics allow SAS to offer solutions to potential performance concerns.
Conclusion
In summary, end-user performance of SAS Drug Development is a highly complex issue to address and includes many factors to consider. Analyzing performance metrics through the methods described above allows SAS to generate a baseline for customers. This baseline can be used to grade a customer’s Internet performance. In addition, once the baseline is established, any Internet configuration changes made at a customer site can be checked to determine whether performance issues have been introduced. If a performance issue is discovered, SAS can aid in the investigation.
Operating System and Release Information
| Product Family | Product | System | Product Release | SAS Release | ||
| Reported | Fixed* | Reported | Fixed* | |||
| SAS System | SAS Drug Development - SAS Hosted Edition | Microsoft Windows Server 2003 Datacenter Edition | 3.5 | 9.2 TS2M3 | ||
| Microsoft Windows Server 2003 Enterprise Edition | 3.5 | 9.2 TS2M3 | ||||
| Microsoft Windows Server 2003 Standard Edition | 3.5 | 9.2 TS2M3 | ||||
| Microsoft Windows Server 2003 for x64 | 3.5 | 9.2 TS2M3 | ||||
| Microsoft Windows Server 2008 | 3.5 | 9.2 TS2M3 | ||||
| Microsoft Windows Server 2008 for x64 | 3.5 | 9.2 TS2M3 | ||||
| Microsoft Windows XP Professional | 3.5 | 9.2 TS2M3 | ||||
| Windows 7 Enterprise 32 bit | 3.5 | 9.2 TS2M3 | ||||
| Windows 7 Enterprise x64 | 3.5 | 9.2 TS2M3 | ||||
| Windows 7 Home Premium 32 bit | 3.5 | 9.2 TS2M3 | ||||
| Windows 7 Home Premium x64 | 3.5 | 9.2 TS2M3 | ||||
| Windows 7 Professional 32 bit | 3.5 | 9.2 TS2M3 | ||||
| Windows 7 Professional x64 | 3.5 | 9.2 TS2M3 | ||||
| Windows 7 Ultimate 32 bit | 3.5 | 9.2 TS2M3 | ||||
| Windows 7 Ultimate x64 | 3.5 | 9.2 TS2M3 | ||||
| Windows Vista | 3.5 | 9.2 TS2M3 | ||||
| Windows Vista for x64 | 3.5 | 9.2 TS2M3 | ||||
| 64-bit Enabled Solaris | 3.5 | 9.2 TS2M3 | ||||
| HP-UX IPF | 3.5 | 9.2 TS2M3 | ||||
| Microsoft Windows 2000 Advanced Server | 3.4 | 9.1 TS1M3 SP4 | ||||
| Microsoft Windows 2000 Datacenter Server | 3.4 | 9.1 TS1M3 SP4 | ||||
| Microsoft Windows 2000 Server | 3.4 | 9.1 TS1M3 SP4 | ||||
| Microsoft Windows 2000 Professional | 3.4 | 9.1 TS1M3 SP4 | ||||
| Microsoft Windows NT Workstation | 3.4 | 9.1 TS1M3 SP4 | ||||
| 64-bit Enabled HP-UX | 3.4 | 9.1 TS1M3 SP4 | ||||
| Type: | Usage Note |
| Priority: |
| Date Modified: | 2012-01-12 10:38:42 |
| Date Created: | 2011-12-05 15:15:59 |