Latency always has been an issue for users of the internet, but response time arguably is becoming the bigger problem. In a strict sense, “latency” refers to transmission network issues.
Response time is the amount of time a computing system takes to react to a request, once it has received one. The time between invoking an application programming interface and the instant this API returns the result of its computation is an example of response time.
In the era of cloud computing, when the servers are located remotely, issues such as broker service policy, load balancing technique and scheduling scheduling algorithm affect cloud data center response time. But there are multiple sources of potential response time lag.
A study by Decibel suggests latency and response time still are issues. Conversely, web pages that load fast are at the top of the list of attributes of a website that contribute to a positive user experience.
Over time, the amount of network-induced latency has been dropping, for a number of reasons, including use of content delivery networks that cache content at the network edge. But optical wide area networks these days minimize latency, and each mobile network generation also has featured lower latency.
The 5G network will essentially eliminate access network latency as an issue for user experience. But that only shifts the latency problem to application server response, which is why edge computing, with its promise of lower latency, now is becoming an issue.
As network congestion and bandwidth become less problematic, end user experienced latency increasingly becomes a matter of far-end application server response time.
In the past, propagation delay, or delays caused by transmission networks, arguably have been a chief cause of user-experienced delay.
A basic rule of thumb for web page sessions has been that 0.1 second is about the limit for having the user feel that the system is reacting instantaneously.
Latency of about one second is about the limit for the user's flow of thought to stay uninterrupted, even though the user will notice the delay.
Some argue that 10 seconds is about the limit for keeping the user's attention focused on the session.
But it is gaming and other coming applications where latency really gets tested. Though the solution to user experience issues often includes an increase in delivered bandwidth and a reduction in transmission congestion, that does not solve all latency issues.
Observers often focus on network latency, or ping time, the time it takes in milliseconds for your network to connect to the internet host and start uploading or downloading data. That matters, of course.
Most 4G LTE connections have featured an average latency of less than 70 milliseconds, which is far below the recommended maximum latency of 150 milliseconds for Xbox Live.
For many use cases, the time it takes a server to respond now becomes the chief latency culprit, however.
According to a study by Ashraf Zia, M.N.A. Khan Department of Computing and Shaheed of the Zulfikar Ali Bhutto Institute of Science & Technology, Islamabad, Pakistan, when the user base and data centers are located in the same region, the average overall response time is 50.35 milliseconds.
When the user base and data centers are located in different regions, the response time increases significantly, to an average of 401.72 milliseconds. That study probably reflects transmission latency more than cloud data center performance, however.
Still, one advantage of edge computing is that it removes most of the indeterminacy of transport across wide area networks.
No comments:
Post a Comment