Packet Losses Reported in Pakistan; U.S. Web Performance Normal
📅 - Matrix.net has reported that since early morning Oct. 8, Internet transmissions from Pakistan have experienced a great increase in packet loss. Almost simultaneously, there has been a decrease in latency for Internet traffic in Kazakhstan. Moreover, a large number of sites in Saudi Arabia have become unreachable.
These anomalies occurred at the time of the second military strike onAfghanistan.
In all likelihood, the packet losses are the result of transmissionsswitching to a phone network. The lack of contact with Saudi Arabia may beintentional: either preserving Internet silence for security reasons, ormerely refusing network pings.
The increase in speed (latency decrease) for Kazakhstan is very likely theresult of military operations personnel establishing faster links (onnursat.net) once the operation had begun.
In related news, measurements of Internet and Web site performance in NorthAmerica show no unusual effects either so far in the 24 hours following theescalation of the U.S. response to terrorism, according to Keynote'sInternet Health Report (internethealthreport.com).
The Keynote Business 40 Internet Performance Index yesterday morningaveraged 2.21 seconds at 8:00am Eastern and 2.73 seconds at 11:00am, atwhich time availability was 90.10%, all entirely within normal parameters.At 1:00pm, availability dropped to 86.68%, which is a little low, andperformance improved to 2.46 seconds. The overall average for the Index fromSept. 24-28 was 2.50 seconds, and availability is typically 90-92%.
The CNN.com site has been performing at under 2.0 seconds this morning, butexperiencing a lot of content errors, meaning that some components of thepage (banner ads, graphics, etc.) are not completely downloading, whichdon't appear to be impacting the ability of users to access or navigate thesite. On Sunday, 10/7, performance was under 4.0 seconds, and under 7.0seconds on Saturday.
Although Keynote also measures such things as packet loss and serverperformance, of interest and importance to IT staff and Web site operators,in and of itself packet loss does not necessarily impact performance asexperienced by users. The Internet drops packets all the time. But routersdropping packets will request retransmission of the data, and this mayhappen so quickly that it's unnoticeable to end users. The effect of packetloss on the performance of a particular page download depends on the page,as well as possible congestion affecting that part of the Internet.
Generally, the bigger the page (the more graphics it has, for example), thelarger the lost packets and the greater the amount that needs to beretransmitted. Some pages can tolerate more packet loss than others withoutnoticeable performance degradation. This does not hold true, however, forstreaming content, which is dependent on the successful transmission ofpackets in succession. The only way to accurately measure performance asexperienced by end users is to measure actual download time as usersexperience it.
These anomalies occurred at the time of the second military strike onAfghanistan.
In all likelihood, the packet losses are the result of transmissionsswitching to a phone network. The lack of contact with Saudi Arabia may beintentional: either preserving Internet silence for security reasons, ormerely refusing network pings.
The increase in speed (latency decrease) for Kazakhstan is very likely theresult of military operations personnel establishing faster links (onnursat.net) once the operation had begun.
In related news, measurements of Internet and Web site performance in NorthAmerica show no unusual effects either so far in the 24 hours following theescalation of the U.S. response to terrorism, according to Keynote'sInternet Health Report (internethealthreport.com).
The Keynote Business 40 Internet Performance Index yesterday morningaveraged 2.21 seconds at 8:00am Eastern and 2.73 seconds at 11:00am, atwhich time availability was 90.10%, all entirely within normal parameters.At 1:00pm, availability dropped to 86.68%, which is a little low, andperformance improved to 2.46 seconds. The overall average for the Index fromSept. 24-28 was 2.50 seconds, and availability is typically 90-92%.
The CNN.com site has been performing at under 2.0 seconds this morning, butexperiencing a lot of content errors, meaning that some components of thepage (banner ads, graphics, etc.) are not completely downloading, whichdon't appear to be impacting the ability of users to access or navigate thesite. On Sunday, 10/7, performance was under 4.0 seconds, and under 7.0seconds on Saturday.
Although Keynote also measures such things as packet loss and serverperformance, of interest and importance to IT staff and Web site operators,in and of itself packet loss does not necessarily impact performance asexperienced by users. The Internet drops packets all the time. But routersdropping packets will request retransmission of the data, and this mayhappen so quickly that it's unnoticeable to end users. The effect of packetloss on the performance of a particular page download depends on the page,as well as possible congestion affecting that part of the Internet.
Generally, the bigger the page (the more graphics it has, for example), thelarger the lost packets and the greater the amount that needs to beretransmitted. Some pages can tolerate more packet loss than others withoutnoticeable performance degradation. This does not hold true, however, forstreaming content, which is dependent on the successful transmission ofpackets in succession. The only way to accurately measure performance asexperienced by end users is to measure actual download time as usersexperience it.
Reads: 3972 | Category: General | Source: TheWHIR : Web Host Industry Reviews
URL source: http://www.thewhir.com/marketwatch/pakistan109.cfm
Want to add a website news or press release ? Just do it, it's free! Use add web hosting news!
📅 -