Cloud Computing Resources for Small and Midsize Businesses

Cloud Computing for SMBs

Subscribe to Cloud Computing for SMBs: eMailAlertsEmail Alerts newslettersWeekly Newsletters
Get Cloud Computing for SMBs: homepageHomepage mobileMobile rssRSS facebookFacebook twitterTwitter linkedinLinkedIn


SMB Cloud Authors: Pat Romanski, Yeshim Deniz, Harry Trott, Breaking News, Breaking New

Related Topics: Cloud Computing, Cloud Computing for SMBs

Cloud Computing: Article

The Big Latency Issues in the Cloud - Can You Solve It?

We think we know what network latency is, but do we actually know it?

The meritorious segments of Cloud computing allow users to avail or rent-out computing and storage resources on-demand to run their various IT related operations and networking applications. This whole process does not even require users to maintain on-premise infrastructure as these resources are virtualized via multiplex of virtual servers belonging to a set of Data centers. These data centers are employed by Cloud computing providers that involve extensive transitioning of data across devices used by end-users.

As virtual data distributing servers that run disparate applications may share a singular data center, in this article, I will empirically evaluate and determine the efficiency of the Cloud infrastructure for running and virtualizing latency-sensitive applications. I will also focus on the varying background loads of the Cloud that tend to interfere with the performance of dynamic latency-sensitive tasks.

The latency jargon for moving applications on the Cloud
The goal for developing any type of application is always to ensure the best of its performance, however, in terms of Cloud applications, or while deploying your applications in the Cloud, there are a lot of subtleties that users need to understand. Often disregarded, these details can impact the latency of your Cloud application in many ways.

Before we would go any further, let's first understand what latency really means while deploying applications in the Cloud and how developers and/or IT leaders can keep a better track of it.

Understanding latency in the Cloud
In the past it was quite easy to understand latency, you only needed to determine the router hops between you and your application and evaluate the delays it took data packets to travel from one location to the other. It was also common for businesses to own all the intervening routers and thus, the delays were consistent and fairly predictable.

However, now and in terms of a Cloud Applications, calculating latency is not that easy. This is because the endpoints for your applications are not fixed. Furthermore, your users can be anywhere in the world and might be using a high-speed fiber line or a satellite up-link for connectivity.

Secondly, the beauty and freedom of the Cloud also means that your application itself can be situated anywhere in the world. And which is why there is a price to that freedom too, the resulting latencies of your application to and fro can be horrific and great.

How latency revolves in the Cloud- Distributed Computing
As mentioned earlier, back in the days when everything was managed via on-premise data centers, it was in a way easier to locate and remove bottlenecks and reduce latency. However, now with the inception of Cloud Computing and Big Data the nature of latency rendering factors are changing.

With Cloud's Grid of servers, the Cloud apps employ Hundreds of servers and compute resources that can be located anywhere in the world. Moreover, this multiplex will have its own varying degree of latency depending on the type of internet connection these servers use.

Virtualization also plays its part in causing latency
Unlike conventional rack mounted servers, today's datacenters are made up of a web of Hypervisors that run dozens of virtual machines. This here adds another layer of complexity for data transitioning, as a virtualized infrastructure may introduce its own delays way before any data leaves the rack.

Tracking down CSPs infrastructure

The final most complicating factor that causes latency in the Cloud is the un-standardized or un-orchestrated infrastructure that your Cloud provider employs. However, some Cloud providers are cagey in narrating their configurations. Let alone connections and gear, some CSPs do not even tell you about their data center location. Anyways, if your CSP is on the down side, reason to your latency pains can be none other than them.

How to reduce latency in the Cloud
Now that you have a better understanding of some of the factors that increase delays, the next step is how you can address this issue and reduce latency in the certain segments of Cloud computing.

Here, we would also like to add a whitepaper from Arista Networks that mentions some broader areas users can focus on to reduce latency.

We believe that Arista puts some great advices to be used for reducing data or network switch transit time, however, it is still important for users to examine other more mundane segments of their CSP's offerings.

Address your CSPs
In order to determine possible latency related issue, it's a foremost job of every developer or business owner to see if their Cloud service provider employs standardized IT infrastructures. Furthermore, there are a few areas which CSPs should highlight to a user either in their SLAs or policies. Such particular questions that users need to ask are.

  • Are the virtual machines settled in different hypervisors?
  • Will the user have any say in processes that tend to have an impact on latency?
  • How many router hops will be there in your Network?
  • What bandwidth will be used in their own infrastructure?

Content delivery networks can also help
Content delivery networks are renowned in the whole world for super charging the delivery of web content to distributed end users. However, as they are mostly used for delivering static content they may not work that well for many Cloud applications. Cloud applications require a different type of treatment, and if you are not using static pieces of content, there is not much you can do to improve your latency through a CDN.

Nevertheless, CDNS are enhancing their capabilities by caching more than just static HTML pages. Certainly this fact can get you going and may help to reduce your latencies; however, this matter needs a bit of investigation on whether a CDN can solve your particular situation.

Dedicated or VPS Hosting options
Dedicated hosting and virtual private hosting may have their offerings placed in a rather conventional and inflexible manner, however, experts regard that the "Sole-Owner" methodology of these two can grant you with optimized online experience parallel with reduced latencies.

Still, hosting options such as this may prove to bottleneck your uptime, with issues of limited resources and vulnerability. The decision is yours, but it's always sensible to first address different hosting providers and determine whether they comply with your requirements.

One important tool for measuring Cloud latency
To better understand latency issues of your Cloud application, Compuware has created the smart tool i.e. CloudSleuth. This tool lets you mirror the real world conditions of your app by simultaneously accessing it form 32 different locations around the world. This tool then lets you measure the response time of your app hosted on 17 renowned Cloud service providers.

Over to the developers
It also relies on developers to inhibit the latency related crashes and burnouts of their applications. Even if the latency is there, they can mend the way their applications tackle with this problem by designing them smartly.

Here, they need to understand that creating unwanted back-and-forth computer communications and sloppy coding decisions encourages long latencies and fail-over's in running their applications.

Conclusion
I agree to the fact that Cloud latencies are not just about the trace routes and router hops residing in the IT infrastructure of CSPs'. This issue has several in-depth layers and dimensions with varying complicating factors. Hence, it is imperative for developers to examine each of the factors thoroughly and always take their Cloud providers in the loop to better understand the causes.

More Stories By Sabih Ahmed

Sabih Ahmed is a digital marketing professional and co-editor at Cloudreviews. He has been associated with #VPN and #Cloud computing industry from more than 2 years. Follow him on Twitter @ssabihahmed