Richard's Blog

Really good blog from J P Rangaswami about knowledge and how spreading it wider is a good thing. You should read it. Actually I’d recommend checking his blog regularly – some thoughtful and smart posts on IT and the internet in general. Aside from being a fine comment on the effect of the Internet on society, this post aligns with a basic principle at NetEvidence – that information about the state and health of networks and applications should not be confined to the engineering-minded few, but available to everyone, on both sides of the ISP–Enterprise relationship. Business people. Sales. Service. Finance. Management. The comments we get back from people (and a 50% growth in subscriptions over the past year) illustrate that JP is right, and that this approach works.

Does anyone get what IBM are doing in terms of Performance Management these days ? I talked to an ISP who’d just bought their ‘Proviso’ tool the other day, but turns out it was ‘slipped in’ as part of a larger deal and they hadn’t really chosen it independently. As a product IBM don’t seem to be doing anything with the Proviso component of Tivoli since acquiring it when they bought Micromuse in 2005. It’s generally regarded as clunky and old, it isn’t SaaS, and it really isn’t in the right place to be supporting IBM’s drive towards SaaS (which in contrast seems pretty well thought out but then, if you want the world to buy more servers, preferably IBM servers, SaaS is a pretty logical direction to be pushing). Would love to hear from anyone who can tell me what’s going on there.

You’d think Apple’s marketing had reached everyone, but apparently not. It was late-morning one day last week and my wife came home to find that the cleaner, who comes in once a month to tidy around, had mistaken our iPad for a large coaster, and had rested her large wet bucket of soapy water on it while she washed some windows. I wish I’d had a video of Julie’s reaction, but I think the lady in question now understands Apple’s baby, at least in general terms. Brownie points to Apple for making it at least moderately waterproof.

We just ran some quarterly reports and found that our subscriptions count – one of the key metrics we measure our success by – is up by 50% over this time last year. That’s great news and a vote of confidence in the team here. It coincides with our best-ever month, and best-ever quarter, in revenue terms.

To read the press release, download the pdf  NetEvidence’s SaaS subscriptions increase by over 50 per cent

Good to find small tools which make your life easier. I have never got on with the ‘To Do’ list in Outlook, but this one I love – you can manage any number of lists and structure your projects in a ‘tree’. If you’re a list-driven person, take a look.

Just heard we’ve won a very cool contract with a major retailer – through a partner, so can’t say who it is, but they’re doing some very interesting work with multimedia / HD video and want to be sure their network will cope while running the rest of their key applications. We beat some sizeable competition to win the business – but won specifically because the customer preferred NetEvidence’s clear, usable reporting. Looking forward to watching this one roll out although the customer wants it yesterday and we have to keep pointing out that SaaS is quick to deploy, but not that quick.

Speaking to a number of Mobile operators recently about the effect mobile data has had on their network. Not ‘news’ of course, but interesting to hear their side of it.

We’ve seen mobile carriers increasingly specifying IP and Carrier Ethernet for their backbone upgrades, but people usually cite Cost as the reason, and overlook Flexibility. These guys have had their fingers burnt by building backbones in the early part of this decade from ‘legacy’ E1 or similar TDM lego bricks, and ended up with a network which is hard to scale in capacity terms. 

This sounded OK in the early days of smartPhones because things changed slowly. Blackberries just shunted emails back and forth, and when the iPhone was launched, many assumed it would have a  similar effect. Very wrong. The basic iPhone, of course, produced a false sense of security because it was just that – a phone. It took Apple’s launch of the 3G version, in late 2008, to turn on the firehose. iPhone sales shot from 700,000 in Q3, to seven million in Q4, and users never looked back. One of our guiding principles at NetEvidence is that providing a good user interface is key – more important, in fact than sheer depth of data. The iPhone illustrates this nicely : simply dump the web on a phone, as people had done previously, and little happens. But make the web usable on a phone, and the world tilts on its axis. Stuck with old-style TDM backbones, operators saw their traffic growing exponentially in a matter of months, but lacked the flexibility to do much about it. We’ve all seen the result. Carrier Ethernet – plus the right core hardware from vendors such as Juniper who 'get' this space -  allows them to turn bandwidth up and down easily. Assuming of course they have a good performance reporting solution in place so they know when to change things. Which reminds me…

You know you’ve done something right when you get user feedback within hours of deploying a new software release.

“Re the new version : I’ve been pulling the reports this morning – the improvements are amazing !”.

Significant applause to the Development Team, I think.

Yesterday I’m in the hardware store buying some tools to try and start a chainsaw that’s lain unused for two years. And I grab a couple of small screwdrivers because I can never find one and I realise there are certain fundamental items that – no matter how many you buy – are never there when you want one. I figure there’s some kind of quantum effect here. The universe will maintain a level of these items just below what’s needed and anything you buy that exceeds this level will spontaneously disappear. Tape measures. Staplers. Radiator bleed keys. There’s a research project there, somewhere.

So you’re selling a classic, on-premise application, and your marketing team insist you need a SaaS version, and fast. How do you do it ? By buying up some big servers in a hosting facility and running a separate instance of your software for each customer. Seriously. A separate copy of the software for each client, with the obvious problems of scalability and manageability (like when you’ve got 300 customers and need to roll out a bug fix) and limitations in potential for sharing and communicating. This is called Single Tenancy and it's frustrating because when these suppliers and their customers hit these problems, SaaS will undeservedly take a kicking. But I've seen a number of vendors doing it recently, from very small to very large. If you're considering using SaaS for any application, make sure you understand this Single- versus Multi-tenancy aspect because it gives you a good clue how SaaS-ready the vendor is. Phil Wainwright's recent article is a good place to start. Or talk to us.

Rolling out a new release of Highlight today – with yet more scalability and performance (alongside VoIP and VLAN features – but that’s another story). Performance has always been key for us; we’re handling large volumes of data and we want Highlight to be really fast and responsive, encouraging people to use and explore the application. A long time back we came to the conclusion that classic, transactional SQL databases simply weren’t fast enough to get the speed we wanted. We’ve spent a lot of time developing our own high-speed database technology for network-collected data, and I was pleased to see at the recent TM-World exhibition that the analytics industry is starting to catch on, with at least one company – infoBright - selling specialised databases for performance data. We do use SQL for administrative data, layered with some cool caching technology, but Highlight is proof that SQL isn’t the right tool for every job.

Interesting conversation yesterday, as part of some product-development research, with someone in the Low Latency team at a major London investment bank. They have some serious monitoring requirements, although the most critical systems still short-cut the network altogether by being co-located in the same data centre as the information provider they’re dealing with. Breaking through that barrier of trust will be a major achievement for service providers. My favourite comment though : “We will be looking to Service Providers to give us more visibility into what they’re providing. This will become a differentiating factor. We will be more demanding in terms of reporting, we’re going to be watching the level of service delivered and if we can’t see enough, we’ll just move away”. Couldn’t be any clearer.

We’re in an old building and there’s always been a real antique computer sitting in one corner : a DEC PDP-8. It was there when we arrived and we kind of inherited it. Despite the fact that there’s more processing power in the average mobile phone, and we’ve never so much as powered it up, it had a reassuring solidity. Until last week that is, when we finally decided we needed the space, and had to find a new home for it. Incredibly, we found a guy about ten miles up the road who collects them (that’s one very understanding wife) and he took it away 48 hours later. Best of luck, Toby

Just finished a briefing with Bojan Simic, chief anaylst at TRAC Research in Boston. We’ve been talking to Bojan for a couple of years now and he’s writing some great stuff on Performance Management for applications and networks. Apart from an update on our roadmap, we talked about the tendency of the larger Service Providers to have made extensive investments in older monitoring platforms, and now find themselves struggling to keep up with developments in the market. We’re seeing two trends illustrating this. First, those providers are finding it increasingly hard to respond to the demands Enterprises are making for SLAs in new contracts  – they don’t have the flexibility to provide bespoke SLAs for different customers, for example, or to report around newer metrics like MOS on Voice. And second, those same frustrated Enterprises are going out and equipping themselves with newer, more agile reporting tools (like ours, I’m  pleased to say) and then dictating to the MSP how things are going to be. Those MSPs are losing customer confidence, and customer stickiness, and actually reversing the outsourcing trend they’re working so hard to promote. Not good.

A certain volcano is quietening down, and the skies are back normal. Really sorry for those whose travel got disrupted – but for the first time in my lifetime, the skies were empty. Really, totally empty. No vapor trails. No gentle but constant background jet noise. Just birds. Couldn’t stop looking up and enjoying the peace.

Nice article from ZDNet. Plenty of traditional software companies (including some in our sector) are claiming to have SaaS options. They (and their customers) need to go carefully.

http://blogs.zdnet.com/sommer/?p=831

We’re 100% SaaS and have been for eight years. If I’d written this it would sound biased. But Brian Sommer has nailed it here, I think.

I love the fact that our service is so visual. I’ve done two demonstrations recently by way of helping out our Sales team and it confirmed to me that our User Interface really works. A good UI is so important - look at how Apple have built a global brand based on that principle. When we do competitor reviews, particularly looking at software from the ‘big players’, I’m amazed at the density of information they’re asking users to take in.

Last week, demonstrating to a prospect in the City of London, we got complete involvement from the customer and a comment afterwards of "This is streets ahead of anything else we’ve been shown". Demo number two included people with pure Financial Management backgrounds. “I really, totally got that” said one of them afterwards. “I honestly thought beforehand, I’m gonna be asleep within five minutes but I found myself thinking, I actually understand this”. Wow, I thought, all that effort in the interface pays off. Thanks for the feedback, people!

Our customers requested MOS Reporting to help plan Voice-over-IP rollouts, and monitor qualilty afterwards (some have even got MOS SLA targets with their Service Provider). And it's a surprise to find a metric that gets good as it gets higher. Big numbers are bad in almost everything you measure - traffic, delay, jitter, response time... Not so with MOS: 1 is bad, 5 is best. So we've added MOS Scoring to Highlight's Performance Insight feature, ugh I have to stop saying that as I keep being reminded that MOS expands to Mean Opinion Score so the extra 'scoring' is redundant. Pedants. I know that.

We just closed another quarter – and our finance people (you have to love finance people who get you a quarter’s results just two days after it closes) tell me it’s a record one for us. Despite a challenging financial environment we grew 25% on the same quarter last year.  Revenue is better-balanced across our partners, and we’re seeing more people interested in multi-year contracts, which has to be a good sign.

Net Evidence continues to grow and so we're recruiting again. Growing the Development team this time, and looking for a strong C# Web developer – you can find the jobspec on our website, or mail your CV to careers@net-evidence.com.