I have to do a lot of tech support for my extended family and it's tedious. I wish I could put them all onto a stable, decent operating system that didn't keep patching itself, downloading virus updates and requiring resource slurping firewalls...... let's not go there again.
Particularly annoyed with Tesco this time. My stepfather's (my kids call him Grandpa) email stopped working and so he started an email conversation with the support line on my mum's pc. I was out of contact at the time so he bravely went ahead, what could be the worst that could happen after all? ;-) Now he's nearly 80, so you can excuse him for not knowing everything about how the internet works.
The support people (incorrectly) surmised the problem was with his antivirus software (AVG from Grisoft.com - very good btw). So they asked him to uninstall it and try the email again. Fair enough. The email didn't work. So they told him to reinstall AVG...
"How do I do that then?" asks Grandpa.
"Go to a search engine and type AVG" says mister Tesco support.
"I've got 'Free Antivirus downloads'. Is that the one I want?" asks Grandpa.
"Yes, that's it. Download that and run it." says mister Tesco support.
Well, guess what. If you type AVG into askjeeves, the first entry that comes up is a sponsored link. It does indeed say 'Free Antivirus downloads", but it sure as hell isn't AVG from Grisoft. It's some dodgy outfit called stop-sign.com and a quick search on google reveals it's a nice bunch of spyware. When I found out I told Grandpa to pull his network card out until I got there. Nice one Tesco.
Bearing in mind that this 'tech support line' costs 50p/minute, I think we deserve someone a bit more savvy on the end of the phone. Hardly surprising we end up with this is it?
As we are wont to do in the UK, a stiffly worded letter will be sent.
( Oct 21 2005, 10:02:12 AM BST / Oct 20 2005, 05:42:38 PM BST )
PermalinkComments [0]
Trackback: http://blogs.sun.com/Drew/entry/tesco_killed_grandpa_s_pc
Wednesday October 19, 2005
A little bit more on Digital Certs...
Just a quick update as I'm off on holiday soon and a tonne of work still to do...
I'm definitely in the wrong part of the Identity Profession, by the way. Robin Wilton, my corporate counterpart, got to spend a sojourn in Singapore of all places, whereas I got back ache during stand duty at SOCITM in Brighton. Yes, it was grey and yes it was raining. Do I feel sorry for him cooped up in an economy class seat? ... only a bit. ;-)
Anyway, following on from my last entry.
The ID Card survived its third reading yesterday. Quite a large amount of opposition on the back benches there. On to the rigourous scrutiny of the House of Lords. ;-)
It also appears that this National ID Card and/or Digital Certificate stuff is still all over the place. I heard today that Mr. Clarke fully intends the card to be capable of carrying a certificate - I'll try to get a link for that one soon. Does that mean we're going to have to pay extra for our own Certificates on top of the basic card? That won't help the digital divide. The rich will be able to afford extra security online, the poor won't. Who knows, but it will be interesting to see what happens next.
Oh, and there was this bit of encouraging news for ministers on how much better multiple biometrics are as opposed to just one.... not!
( Oct 19 2005, 05:02:00 PM BST / Oct 19 2005, 04:42:37 PM BST )
PermalinkComments [0]
Trackback: http://blogs.sun.com/Drew/entry/a_little_bit_more_on
Thursday October 13, 2005
Interlude on Digital Signatures
Geraint's comment on my previous post was rather revealing (thanks for that Geraint). I wasn't aware that the government had specifically ruled out digital certificates for inclusion on the ID card. Hands up, red face. Doh! :-(
This is apparently to reduce the projected cost of the cards... How much does a digital cert cost? About £25 today. If we're going to have to have ID cards, I'd rather pay the extra thanks.
Apparently, other than biometrics, the only other two factor identification component being suggested is the good old PIN number, making the ID Card about as secure as a credit card for online transactions. Not much of a step forward. I am assuming that the Government doesn't intend to issue all 60 million of us with a personal biometric scanner ;-)
Ok, the government has said that the card wasn't designed for e-commerce (real opportunity missed for reduction in fraud, spamming, phishing etc), but it leaves another big problem.
With no digital credential aboard, duplicating a card is going to be, if not easy, not nearly so difficult. It's not easy to guarantee uniqueness without a PKI.
Having thought about it for a while, I can't see a way around it. If you're going to create a system like this you might as well do it properly. I think the government will do a u-turn on this once the detail comes out after the public consultation is over. (What Sir Humphrey called the 'Janet and John Bit' ;-)
Kind of interesting that the government has been looking at ways to make PKI affordable, yet hasn't tied this up with the ID Card. Atleast not yet.
Aren't government departments supposed to talk to each other? ;-)
( Oct 19 2005, 05:05:17 PM BST / Oct 13 2005, 05:05:20 PM BST )
PermalinkComments [2]
Trackback: http://blogs.sun.com/Drew/entry/interlude_on_digital_signatures
The Swizz of the Cards...
Another piece of excellent eclectech animation for you on the subject of ID Cards...
The 'Swizz of the Cards' - No, you're not in Kansas anymore... Lots of fun. ;-)
There are a few of serious points in this that do bear examining though.
The author of this little ditty calls into question how well the registration process for the card will work (ably illustrated by straw man handing over a fistful of readies). He/She/They are correct in pointing out that registration is the point at which fraud is most likely, a problem we're working on right now... sorry can't give out the details at this stage, but you'll like it.
Secondly, there is a glaring mistake on the use of the ID Card to verify somebody elses transaction. In the animation the card is stolen and inserted into a 'PC' to verify a fraudulent transaction. Because it is a valid ID card the transaction goes through.
The author has missed the point that the card needs to be valid and somehow authenticated with the individual using it before it's accepted (two factor authentication). Just pinching somebodies ID card won't work, atleast, not digitally. (Might do as photo ID though). A bit of scare mongering going on there m'thinks.
How is this authentication done? Well biometrics is one answer (and yes, we all know about those problems), but there are other answers too. We're also working on how you might be able to 'switch off'/invalidate a card without the card being physically anywhere near a 'terminal'... can't tell you about that either, but you'll like it. ;-)
The point that you don't need to have multiple identities to commit fraud is a good one though. Most benefit fraud is due to false claim reporting, not multiple identities.
Still, a good piece of animation, looking forward to more!
( Oct 13 2005, 02:22:23 PM BST / Oct 13 2005, 01:37:32 PM BST )
PermalinkComments [1]
Trackback: http://blogs.sun.com/Drew/entry/the_swizz_of_the_cards
ID Cards and the NIR Part 3 - Some thoughts on scale
Scale matters.
Now we get to muck about with some figures. Back of a envelope figures admittedly, but it does allow us to analyse this a little.
There are approximately 60 million people in the UK, and about 3/4 of these are entitled to an Identity Card. So let's say 45 million.
The Government wants to store 51 pieces of personal information on us. This, to me, looks like about 10k's worth of textual data, not counting how much it takes to store biometrics, images and so on. Biometric data seems to vary from hardly anything to 1000's of bytes so lets say another 10k for that. How big is a decent quality jpeg of yours truely? Well I reckon you could get away about 30-50k. Plus we need to add say 20% for encryption wrapping etc. So lets, for sake of easy calculation, say a given user profile starts at 100k per individual. Doesn't sound too bad.
So how much data storage are we talking about? 45,000,000 times 100k = 4,500,000,000k or 4,500,000megs or 4,500Gigabytes. Sounds a lot but it's only 4.5 terabytes and that wont cause much of a problem for a decent storage provider. That will probably fit on my PC in a couple of years time!
What about network bandwidth? This begs the question, how often are you going to use your ID Card? Lets start off working from where we are today. How often do you need to prove your identity to a high level of certainty? Open a bank account? Apply for a credit card? Maybe 3-4 times a year, probably less.
Ok, lets assume the whole card population uses their card 4 times a year. That is 180,000,000 card accesses per year, which is 493,150 per day, which is 20,547 per hour, which is 342 per minute, which is 5.7 per second. A fair old whack, but nothing that will scare the worlds greatest directory product(which, btw, has been benchmarked easily over 10,000 reads per second in a multi server environment, and even a single cpu server can support 1,000 reads per second.)
Yes, the network topography will be interesting, but that atleast is do-able.
How much traffic would be transferred for each access? Tricky, but at the worst case, assuming we transfer the whole 100k profile for each interaction and we need another 100k for protocol and security, that gives us 36 terabytes per year, 100 gigabytes per day, 4 gigabytes per hour, 1 gigabyte a minute and 17 megabytes per second. That's a nice collection of T1 pipes, no big deal.
So, we're home and dry then? Well, not quite. There are two other things we need to consider. The audit trail and the increase in usage of this system.
Lets take a look at the audit. Everytime you use the card the government wants to know. So your profile is going to get bigger. How big is an audit event? Could be anything from bytes to kbytes. If it is kbytes that storage is going to grow by perhaps a 100% per year, maybe even 300-400%. So we could guesstimate storage requirements at 18 terabytes per year, at the 4 uses per citizen per year rate. No worries, might need a couple of extra Sun boxes, and I haven't mentioned backup and fail over and redundancy obviously.
What about the increase in usage of the card over time? Well, lets say we start tieing the card to bank transactions, tube travel, bus and train tickets, congestion charging and toll booths. How often would you be using your card then? I reckon up to 10 times per day might be about on the money, perhaps even higher. What does that look like now?
Well, rather than 4 times a year the citizen is now using the card 3650 times a year. I won't bore you with more long maths, and assuming we're still auditing everything but based on the estimates above we now need...
Reads = 5200 per second
Network Bandwidth = 15 Gigabytes per second (You'd need 2,666 T3 lines working flat out)
1 Years Audit = 1642.5 Terabytes (assuming an audit event is 10kbytes)
What makes these figures scary? The combination of an audit and the scale. Can't help thinking that this audit is going to be a little on the expensive side.
Hmmm. Just as well we bought StorageTek. Time to buy some shares in Cisco too! ;-)
( Oct 13 2005, 12:31:46 PM BST / Oct 13 2005, 11:13:03 AM BST )
PermalinkComments [0]
Trackback: http://blogs.sun.com/Drew/entry/id_cards_and_the_nir2
Hope he doesn't mind...
Pinched a few ideas from Robin's blog... I thought I ought to get around to adding a bit of 'sizzle' to this blog.. Copied his stat counter and clustermap ideas. Nice one. I also found this neat little moon phase applet.
As one of my previous colleagues used to say - "Plagarise, plagarise, let no ones work evade your eyes..."
By the way, Robin is one of our top Identity heavyweights. He's the chap who I ask when I get stuck. He also got selected as 'C3PO' in one of our internal, "So which Star Wars character are you then?" emails. :-)
( Oct 13 2005, 09:47:29 AM BST / Oct 13 2005, 09:21:10 AM BST )
PermalinkComments [0]
Trackback: http://blogs.sun.com/Drew/entry/hope_he_doesn_t_mind
Do you suffer from constant tiredness?
Constantly tired, occasionally moody and irritable, generally a little bit unwell a lot of the time but with no obvious cause? Been to the doctor and testing for various things with nothing amiss? You might be suffering from the effects of this little monster:
It's basically a form of yeast which furs up your digestive system and gets all over your body, giving your immune system a constant unrelenting fight and draining your energy. It's a bit controversial, many doctors don't acknowledge it at all (ours hadn't even heard of it)
Check it out for yourself... Symptoms & Causes, and a simple (but a bit gross!) test you can do to find out if you've got it. I had it, and so did my wife - apparently about 80% of the population have it to one degree or another, and it's due to all the crappy processed food we eat.
So what can you do about it? Well, there is a diet which cuts out all yeast (no beer or wine!), all sugar (including fructose from fruit!). Which is a bit drastic.
On the Candida Society Website they do mention somne products which are supposed to help. I took an appropriately skeptical look at these, but my wife wanted to try it out (she's a long term tiredness sufferer) and we thought we'd give the most promising looking one a go - Threelac. I figured it might well be 'snake oil', but it was worth a try.
We both started taking it according to the instructions and to be honest, the first week was awful, the symptoms got a lot worse. This is apparently normal as the 'death' of the yeast dumps a whole bunch of toxins into your system (nice!). As you clear out though, I have to be honest and admit I've felt a lot more energetic. I can get up in the morning without a 'fuzzy' head and my concentration levels having noticeably increased. I passed the test the other morning with no problems. My wife had it worse than me but she is getting better too.
So it worked for us, it might work for you to. Worth checking out if you suffer from unexplained tiredness and generally not feeling 100% for no apparent reason.
( Oct 13 2005, 09:00:35 AM BST / Oct 13 2005, 08:36:54 AM BST )
PermalinkComments [1]
Trackback: http://blogs.sun.com/Drew/entry/do_you_suffer_from_constant
Wednesday October 12, 2005
Things are getting interesting again...
IT hasn't had much to excite joe public in the last couple of years. The browser wars are ancient history, the millenium came without major incident (big round of applause for the much abused IT sector for that - well done people!) and the dot com bubble burst, leaving a few casualties, but contrary to the doom mongers, the internet is still there and e-commerce is bigger than ever...
But where is what we in the UK call, the 'whizz bang' factor?
I think it's coming. Have a look around. You've got people like George Colony, CEO at Forrester saying stuff like "Microsoft is in its most vulnerable moment in history, just like IBM in the 1990s ." That's interesting. Why is that then?
Big threats from open-source and Linux (Yawn, heard that before), big threats from Mozilla (Firefox browser is great, but not exactly life changing lets be honest), big threats from Apple (can anyone say ITunes?)... all cool, but nothing really new there.
What's more interesting is what's happening with the likes of Google, Ebay, Yahoo etc. Our software chief, Jonathon Schwartz (his blog) recently asked one of his audiences the following...
[I] ask[ed] the audience which they'd rather give up - their browser, or all the rest of their desktop apps. (Unanimously, they'd all give up the latter without a blink.)
Speaks volumes that does ;-)
Imagine a world where we can do everything through the browser, office productivity, email, media, collaboration... then imagine the capability of plugging ebay into googleearth, salesforce.com into fedextrack as bespoke 'applications on demand', no need to install, just drag 'n' drop the connections in your browser.
I ran this idea past one of my friends this morning at Starbucks, who in his own words is "a bit of microsofty" who said it would never work, that people wouldn't trust an online system with this kind of data. "You'd never make it secure, and people wouldn't trust it."
My friend said he wanted the option to have his computer disconnected from the network and still be able to do things. I asked him whether or not he wanted to be able to disconnect his house from the mains every so often so he could have fun running his own diesel generator. Ok, a bit flippant, but the point still stands.
The big objection to utility computing is it's not reliable and it's not secure, but how often to you really worry about power outages and weird chemicals in your water supply? You don't, they just work - pretty much. I don't have to know about line voltage, impediance, salinity, ionic catalysts; I just plug in, switch on, turn the tap and down a glass....and pay my monthly bill. My point is these networks are mature, computing isn't yet, but it will be!
What would you need for this brave new world? A basic OS with network and media capability, a reliable broadband connection, a trusted browser, some neat GUI enhancements for that browser, a programming language that works anywhere, loads of online services, storage, security, identity and web services support. No wonder chairs are flying at Microsoft.
Can anyone say 'Sun Microsystems'?
Oh yes, things are getting interesting again!
( Oct 12 2005, 12:30:35 PM BST / Oct 12 2005, 10:25:36 AM BST )
PermalinkComments [0]
Trackback: http://blogs.sun.com/Drew/entry/things_are_getting_interesting_again
Tuesday October 11, 2005
Sxip Identity 2.0 Preso
Far be it from me to promote another company here in my blog ;-), but Sxip Identity's Dick Hardt has made a fantastic preso which bears distribution as a standalone 'Where next for Identity'.
Essential viewing, and fun.
My favourite line is "And usernames and passwords are really cool! They prove you are....
( Oct 12 2005, 12:33:37 PM BST / Oct 11 2005, 05:37:44 PM BST )
PermalinkComments [0]
Trackback: http://blogs.sun.com/Drew/entry/sxip_identity_2_0_preso
ID Cards and the NIR Part 2 - Basic Architectures
Basic Architectures.
Assuming for the time being that we have an ID Card and a Register, what architectures might we consider? Let's take the card itself first. Consider what would need to happen to Identify somebody holding a card. There are four basic alternatives...
1. The card is very simple, containing very little information other than a unique number which acts as a key on the register. All the biometric, data, and personal information would be held on the register.
Pros : Cards very cheap, need not be 'smart'. 'Easy' to perform the audit.
Cons : Everything depends on the network, including the validation of the card and identity check.
2. The card is very complex, containing most or all of the information required, and can provide a local validation capability without resorting to the register. No connection to the register other than at registration time.
Pros : No need for continuous network, cost of running the register much lower
Cons : No audit capability, no fall back if card verification fails
3. The card contains the validation and identity checking capability, but not the personal information. Personal information is held on the register.
Pros : Offline validation and identity checking, some auditing capability for access to data
Cons : Incomplete auditing (can't audit card validation)
4. The card contains all the information. It is a copy of a citizens profile on the register.
Pros : Full Audit, Fall back in case of network failure
Cons : Expensive cards, synchronisation issues between card data and register
Now lets look at the Register. How could that work? Once again there are four basic alternatives.
1. The register is a single large store containing all the required information in one place.
Pros : Simple to administer, easy auditing, conceptually simple
Cons : Severe security issues, severe reliability issues, no privacy controls
2. The register is a distributed database similiar to above, yet with many 'copies' of the register information held in different locations.
Pros : As above, though slightly less so. Better reliability than above
Cons : Same security isses (possibly worse with multiple targets), no privacy controls, data synchronisation problems
3. The register is a distributed database where different distributions hold separate sections of the data profile (Eg. address in one system, entitlements in another) linked by a master index pointing to the sections.
Pros : More secure (multiple attacks required to assemble a complete profile), some privacy controls could be provided
Cons : Tricky to audit (though not impossible), overall system at the mercy of multiple network links
4. The register is a federation of interconnected autonomous stores
Pros : Privacy enabled by design, High degree of security as individual stores would be autonomous, decentralised identity checking (no single point of attack)
Cons : No comprehensive audit possible as the stores are not directly connected, though individual stores will be auditable by their owners.
Federation is a bit of a tricky concept and so I'll explain that in a future article. Now I've introduced the basics, the next article will explain why some of these architectures will have to fall by the wayside when we start to consider the scale of the NIR and ID Card system...
( Oct 11 2005, 03:44:39 PM BST / Oct 11 2005, 01:35:55 PM BST )
PermalinkComments [0]
Trackback: http://blogs.sun.com/Drew/entry/id_cards_and_the_nir1
Monday October 10, 2005
ID Cards and the NIR Part 1 - Introduction
This is a big subject, but it won't come as much of a surprise that Sun has an interest in the ID(Identity) Card and the NIR(National Identity Register) from a technology solutions provider perspective, our expertise in both of these subject areas speaks for itself.
Whilst the greater part of this project will be in the process, training, distribution and production of the Cards, and the operation of the NIR itself, the technology to be employed bears examination, as many issues involved are not immediately straightforward or apparent. Also the technology employed will have a sizable effect on politics, privacy, citizen adoption and security. Where this effects exist, I will point them out.
Thus, I intend to write a series of articles examining the technology based on information in the public domain about the requirements of the ID Card and NIR System, in light of the fact the real requirements don't as yet exist! (Ref6)
-=-=-=-=-=-=-=-=-=-=-=-=-=-
Introduction
Lets examine the Bill(Ref1) itself as a starting point.
There appear to be 4 key areas.
1. An enrollment service, enabling the government to enrole citizens to a high degree of assurance. This includes biometric information; Iris recognition and fingerprints appear to be in vogue at the moment. Enrollment will be in person, i.e. face to face. This is likely to be the most common point of fraud or compromise based on evidence from other ID Card schemes.
2. The construction of a National Identity Register (NIR) which will be a "new, highly secure database holding basic personal information"(Ref1). As No2ID(Ref4) points out, that's "51 pieces of basic personal information" ;-), some of which is current, some historical. Entry to this NIR is subject to the enrolement above. Interesting that the word database is used by everyone involved, it's unlikely to be a 'database' in practice. ;-)
3. A Card, which contains none, some or all of the information on the NIR. The quantity of the data on the card will trade off with the cost of the card. It could range from a simple unique code, all the way to 4megs+ of data on the citizen, including photos and biometrics.
4. A verification service which allows a citizen to connect their card with the NIR. This verification will occur at a number of levels, right from 'Card is Valid' right up to 'Holder is entitled to this service'. This section has major implications for the architecture of the system, we will see this later on.
Alongside this are three other major themes regarding the card and its operation.
1. It's an Identity Card. Well Doh! Importantly, it includes a biometric check capability. The government points out that biometrics are absolutely unique (Ref3), which is true; However it fails to mention that biometric checking does not enjoy quite the same level of confidence. The LSE report(ref2) makes much mention of this. With 30-60 million people registered, even a checking accuracy of 99.9% isn't that great.
2. It's an Entitlement Card. It was infact called an entitlement card up until 2003, but this has been replaced with 'Identity Card' due to pressure from the House of Lords. There has always been a tacit implication that to verify entitlement you must have established identity. Nobody, atleast in print, seems to have realised this is not the case. Buying a train ticket entitles you to travel, yet contains no information about your identity. Likewise, access to the NHS does not actually require you to be identified, merely entitled. A simple point, but drastically important from a privacy perspective. The entitlement and identity information could still be present on the one card, but it doesn't have to be accessed together or associated with each other.
3. It generates an audit trail.
This is an interesting area. Whilst the audit is not written into the bill per se, it is inferred by the inclusion of historical data being stored on the Card/NIR, implying updates to the NIR and/or card. (Eg. Previous addresses, any use of the Card for application, modification or confirmation). The Government appears to require an audit trail of card usage. What they will do with this is a matter of extreme speculation, all the way from nothing at all, to the complete orwellian state.
All of these themes affect the types of architecture that might be presented. In my next article I will analyse these.
Links and references
1. The Bill itself...
http://www.publications.parliament.uk/pa/cm200506/cmbills/049/06049.i-ii.html
2. The LSE Report which caused a ruckus in the press over the price of the 'Card'...
http://www.lse.ac.uk/collections/pressAndInformationOffice/newsAndEvents/archives/2005/IDCard_FinalReport.htm
3. Why you should have an ID Card, courtesy of the Home Office
http://www.homeoffice.gov.uk/passports-and-immigration/id-cards/why-we-need-id-cards/
A bit disappointing really. No detail on 'how' some of the assertions they make will be realised. Also interesting that they file this under 'Passport and Immigration' on their website.
4. Why you shouldn't have an ID Card, courtesy of No2ID
http://www.no2id.com/IDSchemes/whyNot.php
A vocal and well organised group who are "opposed to the government's planned ID card and National Identity Register." Does include some robust arguments worthy of attention and their 'interviews' with various ministers are most amusing.
5. Identity Press Coverage, courtesy of The Register.
http://www.theregister.co.uk/security/identity/
Needs little introduction from me. Well informed journos, now there's a thing!
6. "The very model of a modern labour minister" - A bit of fun, particularly if you like Gilbert and Sullivan; Like me.
http://eclectech.co.uk/clarkeidcards.php
( Oct 10 2005, 04:59:31 PM BST / Oct 10 2005, 03:37:25 PM BST )
PermalinkComments [0]
Trackback: http://blogs.sun.com/Drew/entry/id_cards_and_the_nir