The Future For Solid State Technology On Server Motherboards

from the memories,-like-the-corners-of-my-mind dept

For the past several months, the possibilities of solid state technology have been explored for all sorts of applications. Some observers note that solid state technology inside the server are on the way, but others point out that solid state technology might still be too expensive for typical server applications. As the costs inevitably come down, though, the more important questions might be more along the lines of "where should solid state technology be located to maximize ROI?" or "how can solid state technology be used most effectively?" The use of solid state technology in a variety of new configurations will ultimately depend on the cost effectiveness of the applications, but the potential of 1 million IOPs in the performance of a single server over storage racks filled with thousands of hard drives could be getting a lot harder to pass up.

Reputations Can Be Earned Via B2B Reviews

from the honesty-is-the-best-policy dept

While consumer reviews for all kinds of products are ubiquitous in online forums, the same is not quite true for business-to-business (B2B) reviews. And for the most part, the use of B2B reviews to evaluate products sold to other businesses hasn't become as incredibly widespread. Sure, there are a whole host of 3rd-party firms that will rank and rate products and services at the enterprise level. (eg. Gartner, Forrester, etc.) But those services aggregate multiple opinions in ways that can water down the actual experiences of customers. Some scathing reviews are averaged into the results of a survey, and other glowing endorsements are buried in the numbers on a graph. In many whitepapers, the customer case studies are usually solicited and screened, so while they are reassuring, they lack the more complete transparency of, say, feedback on an eBay seller.

Fortunately, the trend of B2B reviews is growing and being adopted by well-known firms that have figured out that verifiable customer reviews are an important marketing tool. The consumer market teaches retailers that the presence of a few negative reviews actually benefits the overall perception of an online seller. From a few studies, the reputation of a service provider that has dealt with complaints admirably is often better than a competitor that tries to hide complaints or negative reviews. Customers seem to understand that everyone makes mistakes, and that the true test of reliability is how those mistakes are handled. If there's no data on how a company deals with complaints, experienced customers recognize that as a red flag.

Obviously, though, an overwhelming flood of bad experiences is never a good thing. But if that feedback is easily gathered and addressed, then at least there can be a benefit for both customers and their suppliers. Open communication lines can help resolve and improve products more quickly. And providing a venue for customer reviews is a significant first step.

The Key To Useful Application Performance Monitoring -- Service First

from the everything-as-a-service dept

In my experience, the keystone to deploying useful application performance monitoring is to first define and build a hierarchical model of the target application as a service. We have long left a world of stand alone servers and atomic applications, and good performance monitoring solutions should be designed from the ground up to account for each technology involved in the application's delivery.

Naturally the scope of data and events involved in monitoring a modern application's performance as a service can quickly become unwieldy, with information being collated and (hopefully) aggregated from a large number of distinct technologies. To further add to the complexity, the software systems used to target and extract data will also vary from technology to technology. While one might choose to license a single vendor's "monolithic framework" instead of programming solutions internally, the vendor's code to gather data from a router will be distinct from the code used to gather storage information, and so on. Contrary to what marketing claims may be made, there is no "magic button" in any monitoring product that will sort out the relevance of data coming from a subsystem as it relates to an application as it is implemented in the environment.

Defining a common named hierarchy of the application's involved technology upfront enables engineers to systematically "tag" relevant data and events coming from a multitude of sources and correlate them to the appropriate application service (or in some cases... multiple applications). Correlation is the key to all-inclusive monitoring, as information without context is useless.

If the company has not already started using named hierarchies elsewhere in monitoring, it might be a good ideas to develop a basic hierarchical naming standard in the open and let other parties who will (or should) be using the standard provide input to the project. Once an environment starts successfully treating and monitoring one application as a service, there will be a drive from all application owners to do the same for them.

Increasing Redundancy Is Leading To Zero Downtime

from the department-of-redundancy dept

It's nearing the end of the year, so not surprisingly, there's more talk about what's in store for the next year and the near future. Out of these discussions, one of the interesting points predicts that zero downtime is moving to replace "five-nines" of availability. It's not exactly a new concept to plan backups for when components will fail (not if they will), but redundant hardware components are becoming increasingly common as the costs for them decrease. Spare processors can be activated at will. Entire facilities can be replicated in disparate geographic locations. There may always be unexpected downtime, but the costs associated with them are trending towards being more manageable.

Servers In Outer Space -- Are Server-Sats Ready To Launch?

from the latency,-much? dept

Given that California's largest energy utility, PG&E, has invested in energy satellites that would beam energy down from space-based solar electricity generators, it shouldn't be too surprising that servers in space might be just around the corner as well. On the optimistic side, servers in space would have abundant solar energy which wouldn't burn fossil fuels -- unlike traditional datacenters, which are estimated to consume almost 2% of the electrical power generated in the US. But on the practical downside, orbiting communication satellites have a latency of almost a quarter of a second (240ms) for geostationary altitudes. There's not much technology for getting around the speed of light, even if lower orbits are considered. Plus, there are the obvious drawbacks of launching rockets and not being able to upgrade the hardware. So server-sats are a nice idea, but hardly ready for commercial ventures just yet.