|
|
|
5,000 Good, 1,000 Bad
Many organisations depend on their connection to the Internet. We were surprised to find, in a snap survey carried out last month, that one in six of these connections is seriously overloaded. We can say this with confidence because NetEvidence is in a unique position: every few minutes, our servers check health levels and performance for thousands of companies' internet links across the globe (52 countries, at the last count). From a sample of 6,000 participants, we found just over 1,000 had traffic levels, during business hours, which would affect performance for staff or customers. Many of them are being truly hammered, at 90 or 100% loading for hours on end.
This is surprising: these companies are all Highlight users. It's possible they're not checking their Highlight reports regularly; or they know but can't do anything about it ; but our most likely explanation is that they think the connection is coping fine. Why invest the time and money in an upgrade, when users aren't complaining?
But that approach is probably a mistake. Congestion tends to come on slowly, as businesses evolve and traffic patterns change. Over time, Users get used to poor performance, and accept it rather than complaining. Internet protocols have a 'four wheel drive' mode which they slip into when conditions get bad, to ensure traffic gets through - so even though users or customers can feel they're wading through treacle, nothing actually stops. This not healthy!
Our advice to Internet users: check that connection. Use Highlight's News Insight capability to email you when things get heavy (you don't even need to log in) or Application Insight to see which applications are causing all that traffic. Remember bandwidth prices have dropped, so your provider may well offer a faster link for a similar price.
Go on - put a smile on your users' faces.

VoiceFlex Partner
NetEvidence has partnered with Voiceflex, one of the largest Voice-over-IP (VoIP) networks in the UK. Using SIP (Session Initiation Protocol) technology, Voiceflex lets businesses make phone calls using their office DSL connection. The calls pass onto the public phone network like any normal call, but using SIP / DSL means more flexibility, and cost savings.
Voiceflex are not alone in doing this, but the company has recognised that DSL (Broadband) lines can fall short of the quality needed to handle voice traffic; to allay customer concerns, Voiceflex will be providing NetEvidence Highlight with their voice solutions, so their customers have real-time visibility of line quality, and know their voice calls are being handled correctly. As Richard Garner, Sales Director of Data and IP Services for Voiceflex explains, “Our customers want to use new technologies but they expect to have the same consistent and stable telephony services. The transparency offered by Highlight means we can be confident that we are delivering the best possible call quality over the internet. Highlight’s reports will also be available across all our ISP and WAN services ensuring upgrades and quality of service can be anticipated and justified within an organisation.”
Read the full Press Release here...

INDUSTRY NEWS
Analysts Comment
"It is one thing to create the technology to monitor network performance, and quite another to present the data in a format that is meaningful to management. Net-Evidence has achieved this beyond the capability of any of its competitors" says Martin Butler in a recent Technology Brief.
Martin, who founded the well-respected Butler Research Group and then sold it to Datamonitor in 2005, now heads up Martin Butler Research (MBR). The new company focuses "Information Efficiency", providing straightforward and clear recommendations to executive-level clients on how best to reduce business costs and use IT resources effectively. In MBR's unique grid-based ranking system, Highlight scores the highest position - "potentially high output for a low level of input (cost)", as Butler puts it. It's worth visitng MBR's site - not only because it says good things about NetEvidence, but it's also refreshing to find such a promising collection of bite-sized, easy-to-read reports written by an acknowledged expert.

Just Run the Business
"There's a trend here," an IT vendor said to us recently "and I don't like it". She was referring to a current shift in the way businesses look at IT, which is centred on the notion of Software as a Service, or SaaS.
To see what she's worried about, you have to look at the current SaaS phenomenon, but strip away a painful amount of hype, market pressure and press coverage. It's a good thing to appreciate if you want to really understand why people like the idea of SaaS, and it's simply this: businesses just want to get on with things.
Running an IT infrastructure is a sizable overhead, but it's been that way for so long that most businesses have simply become numb to the continual hardware upgrades, software patches, training, staff recruitment, maintenance and other stones in the shoe which come with IT. It's part of the background noise of the business.
To realise that you don't have to do this - that you really can just pay a monthly fee and have the use of an application (not the use of hardware, or a server, but an actual, real piece of business functionality that lets you do your job better) takes some getting used to. But once the notion takes hold, it's wonderful. Underneath the day-to-day hubbub, most managers just want to get on with running their business. Being able to push away all that IT maintenance, and never have to worry about backups, upgrades, or the other stones in your shoes that have been there as long as you can remember, is a huge relief. Of course, you have have confidence in the company providing the SaaS, but that's a short- not a long-term issue.
Do we practice what we preach? Yes. We use SaaS solutions for our CRM and EMail, and this month we've moved our Trouble Ticketing system 'into the cloud' as well. We monitor their performance, and we're picky, but we do it. Not having to worry about those systems lets us focus on what we do - run the Highlight service - more effectively.
From an enterprise perspective, where do lower-level cloud services like Amazon's S3 fit into this? For an enterprise, the answer is probably 'not directly'. S3 provides simple, fast, scalable storage for application developers rather than enterprises themselves - but you should still be excited about them because they enable other companies to build and scale SaaS products quickly and reliably... products that you'll find in the business marketplace now and in the future.

LOOSE CONNECTION
Not strictly related to Data Networks, but worth passing on...
Two loose ones this month.
First, if you know NetEvidence you should know we're very keen on good user interfaces : making information actually usable rather than just dumping it on the user. Here's Pattie Maes at the recent TED (Technology, Entertainment, Design) conference, showing off a very cool user-interface that MIT have been working on. We've no plans to put Highlight onto this type of device yet, but I think we'd have a few takers if we did.
Finally - and we really can't see any way to link this one with Performance Reporting, but it's just fun - spend three minutes finding out about Tweenbots. Tweenbots are a project in New York by art student Kacie Kinzer: Kacie turns small robots loose on the streets of Manhattan with 'help me' messages attached, and tracks what happens to them. Amazingly, most robots survived and got where they were trying to get to: the public, it turns out, are suckers for technology if it's got a smile on its face.
See you next month.
|

OTHER NEWS
Tutorial: WAN Acceleration
Wide Area connections are way slower than local ones. This is pretty much accepted in the industry: even with today's metro-ethernet and long-distance fibres, the number of switches, hops and kilometres involved in moving data from site to site means most applications slow down significantly, or even to a crawl, when the server is one place and the client another.
Using the kind of sweeping generalisation for which our industry is justifiably famous, the usual blame falls on application developers, and their insistence on writing products which rely on fast connections. We would, of course, be substantially wrong in saying this. There's a wide range of applications out there, from web-based things specifically designed to run across slow links; to client-server tools which need something fast but reliable; to file-server products which assume a high-speed and local connection. Frustratingly, it's these latter which organisations insist on trying to run over the WAN. The motivation is obvious : put your file servers in a central data centre where they can be cooled, powered, managed and backed-up easily. Make the users access them over a WAN link because, well, that link is a metro Ethernet which is just like local Ethernet, isn't it? that 8M DSL must be nearly as fast as 10M Ethernet ; and that link is only 2M but it costs more than Ethernet so we've got high expectations.
Except that of course it's not that simple. DSL is often heavily contended; 2M has more latency than you'd think; and before you know it users are complaining about slow response times.
This is where WAN Acceleration (WAAS) comes in. By sitting at either end of a WAN link, and doing clever things with packets for certain applications, WAN Accelerators can have a huge effect on performance. "Clever things" boil down to three in practice:
- Compression, which squeezes data like a real-time 'zip' function. Is effective across all applications, though to varying degrees;
- Caching, which keeps local copies of files or data at the user's end. Can have a huge effect on speed but needs to be written for specific applications;
- Application Acceleration, where specific clever things are done with individual applications, such as locally responding to Exchange email queries or Word Documents. Again, only good for specific applications, and different WAS vendors tend to be good at different ones.
We see lots of organisations investigating this technology, but very cautiously : every installation and every environment will produce different results and benefits. Suppliers report a constant flow of loan units coming and going as people evaluate whether the improvement in their particular network is worth paying for.
Interestingly, some recent surveys in this area - for example from Aberdeen Group - find that most organisations don't know what performance their users currently get ; or what performance their users expect; or sometimes, even what applications are running across their network in the first place - suggesting that for many, leaping into the technology is a bit like splashing out on a new pair of running shoes in the expectation you'll automatically start running faster.
For now, if you're an organisation considering WAS, it seems important to make sure you first understand and benchmark your own environment carefully - otherwise you're adding complexity without really knowing why. Never a good idea. If you want to understand things better, try this review by Dave Newman of NetworkTest: we've known Dave for a long time and he has a great grasp of this and other technolgies.

Work for us...
Following a quarter of strong growth, NetEvidence is recruiting!
Having recently grown our Sales team we're now looking for two people to join our Operations, and Software Development groups.
For full details of the roles, click here...

|