Presentation on Cloud and DCaaS - AIPM

Author: 
John Sheridan - CIO & CISO
Category: 
The Department of Finance Archive

The content on this page and other Finance archive pages is provided to assist research and may contain references to activities or policies that have no current application. See the full archive disclaimer.

In the last couple of months I have given a number of presentations at various forums regarding the use of ICT in the Australian Government.  I thought I would share them here over the coming weeks.

The first is from an Australian Institute of Project Management event held in Canberra in late November. All photo attributions are in the slides.

 

Cloud and DCaaS Presentation

Australian Institute of Project Managment Event, Canberra - November 28, 2012

Audio

Transcript

John:

Now I’m going to do a bit of technological rearrangement just while I’m setting things up and the reason I’m going to do so is that I’d like you to see the way that I’ve created the presentation for tonight.  Firstly, one of the things you’ll notice is I’ve chosen a slightly different title, that isn’t because I’m not going to speak about what I said I was going to speak about, but rather ‘cause I didn’t remember what I called the title when I came to do the presentation, which is a problem when you do a lot of presentations.

I am going to talk to you about the theme of Cloud Computing and where we’re going towards Data Centre as a service.  It’s been an interesting journey for us, and before I get into the journey, I’d just like to talk to you about where we are today, or what occurred earlier this week.  In the budget the Government decided it would do some work designed to address the gender balance on Government boards.  Indeed what it wanted to do was increase the number of women, but get balance across the boards that doesn’t exist at the moment.

Now there are a range of steps in this project, some of them Minister Wong has announced already, others are being worked through, and one of the things we need to do is to collect some details on board members, potential board members.  We’ve got a challenge that we have to have a database in place by the 8th of March to do that, but as sort of hypothetically happens occasionally the Minister’s staff told us a day or so ago that they wanted to do something differently, something sooner than that, and indeed what they wanted to do was get the people who are going to be nominating potential champions… they’re going to be champions that are going to be nominating potential board members, to have a way of writing to them to encourage them to do that really very soon, and we needed a way then to collect the information about these potential candidates for boards.  We had to collect the sort of detail that you’d normally collect about people, name, rank, serial number, those sorts of things, also get their CVs added in as well.

Now the people I had working on this project sort of said well we’ll approach in the historical way, what we’ll do is we’ll get a Microsoft Word document or a Microsoft Excel document and we’ll email this document to the people involved and we’ll get them to fill it out and email it back to us, and we’ll put them all together like that.  And I suggested to them politely that that was probably a little pedestrian in the way we might do it, it would require people to have Microsoft Office software, and as much as we love Microsoft, we know that not everyone has Microsoft Office software, and indeed the experience of the Taxation Office amongst others, is that only having an Office alternative or a Windows alternative sometimes doesn’t cheer people up, and we didn’t think that we wanted to address these potential board members in a way that might sort of upset them.

So I said “Well let’s look at something that’s a bit platform independent, what can we do?”  And there’s a discussion of well should we use Adobe form, and there’s issues around that in terms of licensing and how you do it, how quickly can you put it together, and I said “Well we’ve got this new platform, Data Centre as the service, let’s look and see what we can do,” and of course there’s much rubbing of brows and shaking of heads amongst the project managers, ‘cause you know that project managers always want to deliver things, they want to get it there on time, they want the least number of distractions, and they particularly don’t want senior executives helping them, but we just can’t help it.

And I said to them “Let’s look at what we can do quickly.”  So we looked through the catalogue for Data Centre as a service, and you can do that yourself now, and we found a software as a service provider and I said “Well let’s think about what it is we’re doing.” Essentially we’re doing a customer relationship management arrangement, it’s the sort of data that you collect in CRM, so let’s find a Cloud provider who does CRM, who’s on our panel.  Now because there’s data about this that is related to privacy, we’ve got to manage it accordance with the privacy principles, we don’t need something that’s in the public Cloud necessarily, we need something that’s in either a private Cloud arrangement, or properly secured so we could do that.

So there are 945 services currently on the Data Centre as a service panel, but if you use the filter which, as I said, you can do as well, you can find the provider that provides CRM software, that provides it in a private Cloud arrangement, it happens to be in a quite secure Data Centre, one that is on the Data Centre facilities panel, more on that later, and we’ll do that.  I said “Well let’s ring up this provider and ask them what they can do.”  And there’s still much rubbing of brows and shaking of heads, but we did ring up the provider that happens to be Emantra, completely… I didn’t know they were sponsoring this tonight, so that was just a coincidence, I was going to talk about them anyway, but it’s an even better benefit, “How soon can you do this?”  Well actually we could do it… they could do it overnight, well in fact in an hour essentially, set up what we needed.

For a range of technical reasons on our side, not theirs we’ll actually have it live on Friday.  The cost is very reasonable, I suspect they spent more on sausage rolls tonight, in terms of providing that, and when I go back to the Board for this arrangement and say “Well actually, we’re going to turn this on this weekend, it’s going to be ready to go, it’s going to cost a very small amount including 12 months hosting, indeed it will probably be cheaper than setting up the Microsoft Office web form, document, when you take into account the inevitable changes and the time people would do to approve it and things like that.

This is the value of using the Cloud.  Now I’m going to talk today how we got to the Cloud and where things are going, and it’s not the answer to everything, but I think it’s a good start to realise things are changing in the way that we deliver IT services, because of the Cloud.

Now I’ll take you back in the Data Centre story, back to Sir Peter and his review.  In the Gershon Review delivered in late 2008 said that, amongst other things, the Government could avoid a billion dollars in costs in Data Centres over the next ten to fifteen years if we did some co-ordinated work in the Data Centre area.  Now I’d remind you that a billion dollars’ worth of avoided costs and $3.90 will get you a cappuccino in the Minter Ellison building downstairs.  We’re not talking about savings, although we did a bit of that after Sir Peter had delivered his report as well, but we’re talking about here is not going back to the Government for money.  So we still spent, and we were spending when he did his review some $870 million a year on Data Centres, and we’re still going to spend a lot of money on Data Centres over the next ten to fifteen years, I mean we’re a couple of years into the strategy now.

We’re still going to spend a lot of money in doing that, but the challenge is what can we avoid by doing things smarter?  So that was the task given to us initially.  Essentially the first thing we did was deliver a Data Centre strategy.  Now in order to get to the Data Centre strategy we had to do a range of important things.  First of all we had to make sure we understood the technology and the trends in technology that were arranged, so we paid Gardener to do a survey for us to determine that.  Secondly we had to look at demand from agencies, now I’m going to swear now, so if you’re offended just cover your ears, I used to think that CIOs didn’t wake up in the morning and go “Bugger, I forgot to order a Data Centre,” but actually I don’t think that anymore.  This tends to happen much more than you might think, obviously not in an overnight sense, but the notion that demand isn’t actually thought out well in advanced is occasionally true.  There are challenges with what you might do about that.  But we looked at demand.

We also looked at supply, and what we determined was that there are a range of potential Data Centre facilities.  There were facilities that existed at that time that we could, the target being July 2011, that you could put things into immediately, there were other facilities that could be fitted out, ready about the middle of 2012, and there were other Greenfield facilities that could be built.  So when we went out to the market we said there are these three lots of facilities, we would like you to respond to the tender in that way.  And we’ve got arranged 23, Jeff?

Jeff:    About that.

John: 

Something around that, potential Data Centres around Australia in all the capitals but Adelaide and Darwin at the moment, that can take arrangements.  Sir Peter’s work indicated that there were some 30,000 square metres of Commonwealth Data Centre space at that time, and that unencumbered by any sort of notion of co-ordination, that was likely to expand to 60,000 square metres over the ten to fifteen year period.  If you applied every single control that you could to it, it might shrink to 15,000 square metres, but generally speaking it was clear that we could reduce it, or we could at least hold the expansion and get a better arrangement, and I’ll talk about how we did that in a moment.

So we set up a panel of facilities.  We had some interesting things I think, if you think about how you go around negotiating with vendors in these arrangements there are some interesting lessons that we learnt, and I think vendors learnt as well.  If you can be sufficiently hard-nosed and the Gershon business as usual budget reductions did a lot for hardening up my nose, if you go out and say to them well actually we’re going to negotiate contracts to a certain point.  But when the first contract has been signed, that’s the contract.  And if you want your contract conditions and your arrangements met for your particular arrangement, then you ought to be first.  If you’re not first, then we’re going to be much less interested in varying the conditions as we get on.  Now that’s been a hard line to hold, but we’ve managed to do it generally in that Data Centre facilities area, and we’re indeed continuing to do that as we look at more forms of co-ordinated procurement, and I’ll again touch on that when I get to Data Centre as a service.

The challenge then is what have we done with this?  We’ve currently got I think some 3,500 square metres of data space, using both the interim panel and the current panel arrangements under coverage.  We’ve got one consortia contract signed, two other contracts signed, two other consortia contracts in the making, the Data Centre facility panel allows us to gather up agencies to get to the trigger of 500 square metres, or 500 kilowatts, and put them together, sign a lease with Finance that the vendor still builds the agencies, but we get the sort of benefits of the facilities panel for quite small agencies as a consequence of this work.

So it’s been an interesting growth for us in terms of driving changes.  Now along with the changes in Data Centre facilities, we also set up a Data Centre migration services panel, and this panel allows agencies to, and it’s a mandated panel, allows them if they want to do work moving between Data Centres, designing, planning, decommissioning, commissioning, all those things, it’s a panel of providers that can provide those services.

Again some interesting challenges in doing the work here, how it will work out exactly, how you compare a range of providers when they don’t all offer the same services, and how you settle eventually on which ones you’re going to choose.  Also some interesting work about when you’re doing contract or tender evaluation and you get what looks like quite… you can rank people one to five or one to ten or something like that, and on the face of it, some of you would realise, that can look quite reasonable, except that when you look at the sensitivity of that analysis that you find that the differences between three and nine aren’t very large at all, and by including four you really should include all the ones down to nine, how you look at the methodology for picking which vendors to add to the panel in that arrangement was an interesting challenge for us, both in tender evaluation and then in discussion with our lawyers, because explaining sort of the standard deviation notion to the lawyers and getting that across to them so they could see that, I mean we persuaded them, they saw the benefit of what we were doing, but it was some interesting challenges I think, in delivering this level of service.

Now one of the other outputs of the Data Centre strategy was the need to have optimisation targets, because what we found of course is that you need to understand what you’re running in order to understand how you can change things.  And this has an important impact as it turns out, on the predilection of agencies to take up Cloud computing.

 

Let me just start at the beginning so I can explain why that’s the case.  When we did the first benchmarking work as a consequence of Gershon’s review, the ratio of virtual servers to physical servers was 1.1 to 1 across the Commonwealth.  It’s now 2.3 to 1.  The world-wide average is 1.9 to 1, so we’re ahead of the world-wide average.  The challenge though is what we can do in terms of maximum averages.  Virtualisation companies will say well you can get 15 instances per core, in a quad core processor you could get therefore 60 to 1, depending on the load and those sorts of things, and certainly there are agencies that are delivering ratios, across the agency… average ratios across the agency of greater than 10 to 1, approaching 15 to 1 in some instances, but you can get really solid virtualisation.

Now virtualisation isn’t private Cloud, again more on that later, but what it does tell you is if you’re running strong virtualisation ratios in your agency, the benefits you might get from going to Cloud are going to be less than if you’ve got really crap virtualisation ratios to start with.  And understanding what it is the benefits that we’re going to get from Cloud computing, ties in to what it is our predilection to taking it up.  The Data Centre virtualisation targets, we also have PUE targets, optimisation targets and a range of other things.  We’ve got methodologies that tie into the ICT Sustainment Policy, and we measure all of this in our annual benchmarking work where agencies, depending on their size have to respond to a range of benchmarking questions over time.

So in measuring what it is that we’re doing, now a slight divergence, you’re wondering what the image on the screen is, that of course is Windows 8 running inside parallels on my MacBook Air.  Why did I take that photo today, well it is virtualisation.  But the way that the presentation system that I’m using for this, it’s called Haiku Deck, H-A-I-K-U D-E-C-K, it’s an iPad app that’s free, you can buy some themes for it that cost about 20 bucks or something like that.  When you type in, you’ll see all my slides have a large heading and a small heading, you type in those words, it can search to find pictures that exist for those words and all the other pictures but this one I took from the Web, I think there’s one other, took from the Web.  What it does it find those for you automatically or you can search for particular titles, but it provides the providence of the picture, it can confirms that it’s CCBY Licence, so there’s no copyright issues as long as it attributes them, you’ll see on the other slides I think, if the ratio is right, that there’s a little line that says where the picture came from.

What’s interesting about this is this is way first of all of developing presentations that previously used to take me a lot longer as I tried to do those things, to find the right diagrams.  It also, as you probably have noticed, forces you to… first of all it gives you a picture that tells you what I’m talking about in theory, and it also focuses your attention on me talking and not the small words on the slides, and I think it’s a good discipline to be doing as well.  But why does this work?  Because we’ve got mobile devices, because we’ve got Cloud computing, public Cloud, access to all those sorts of things.

Moving on then, we wanted to get to Cloud Computing.  Now interestingly enough when Sir Peter was doing his work in 2008 there was some discussion about Cloud Computing, but I think it’s true to say that it wasn’t something that was on the front of everybody’s mind as it is at the moment.  So although we knew as soon as we started work on the Data Centre Strategy that we’d have to do something in the Cloud Computing area, it wasn’t new to us when you would do it.  Now I sort of think of a cycle, Gardener talks about the hype cycle, I’ve got my own version of this, firstly researchers talk about things, then vendors talk about them, then providers… sorry, then journalists start talking about them, then Government starts talking about them, and then auditors start talking about them.  And at the moment we’re somewhere between the sort of vendors and journalists talking about them, in most of the Cloud Computing work.

The market… a year ago I think I would have said that the market in Australia wasn’t all that mature.  I wouldn’t say that now, I think it’s much more mature and it’s grown particularly in just 12 months, but there’s no doubt that there’s work required to get us to there, so in 2011, after consultation, quite wide consultation, AGIMO published a Cloud Strategy, and you can still see that Cloud Strategy on our website if you want to have a look at it.

We then started to do some work on the sort of guidelines that would be needed, but in examining the Cloud Strategy we discovered some interesting things.  First of those is that many of the technologies that are used to sell or to make up Cloud Computing, to make it what it is, aren’t in themselves all that new, it’s the combination of those things that creates the Cloud Computing environment. Now you don’t get it just by whacking one of those things on your own shed, the notion that I’ve now got Broadband so my outsourced Data Centre is a private Cloud isn’t true.  The notion that I’ve got virtualisation and therefore my outsourced Data Centre is a private Cloud isn’t true. The notice that you use the internet to get to my resource somehow, that that makes it the Cloud, I think is also not true.  But of course you wouldn’t be surprised to know that many vendors have all sorts of explanations about what they’re doing is in the Cloud and I’ve had some quite amusing discussions with vendors as they’ve come to speak to me to talk about why what they’re doing is Cloud Computing, and it’s always interesting to ask them a few questions around that to find out if it is.

The first of those questions are around the characteristics of their Cloud Computing, now these are the characteristics that are in our Cloud policy documents, they’re very close to those that are in the NIST Cloud Standard run by the US National Institute of Science and Technology, so it’s a relatively well accepted definition.  But first of all is this notion of on-demand self-service, so to be Cloud Computing you’ve got to go and be able to set up what you want by yourself, and that’s the way it’s established by the vendor.  You need broad network access and I think all of us understand why you need that sort of arrangement.  You need it not just in terms of big, fat pipes, but the notion of Cloud Computing is that you can access those resources from a range of areas.  Now many of us would have had the experience of using iTunes and iMacs on our iDevices and realised that this ability to access that data wherever you are is a very important part of the Cloud experience.  Indeed in public Cloud sense perhaps the most important part of that.

The notion of resource pooling is very interesting, and it’s particularly interesting in the context between the public Cloud and the private Cloud.  What we see a lot is that vendors, and I don’t wish to offend vendors, vendors have to make a profit, if they don’t make a profit their companies fail, if their companies fail we won’t get the supplies we need, so I’m just more pointing out the healthy regard I guess I have than anything else, but the challenge for vendors is they talk about public Cloud because there’s no doubt that there are particular cost benefits in public Cloud.  But many of those benefits arise from the amortisation across very large numbers, world-wide amortisation in the case of those big providers.  And it’s important I think to make sure that when people then talk about private Cloud, they’re not using the numbers that sold public Cloud.  Now that’s not to say you don’t need private Cloud, there are reasons of security, privacy, all sorts of things that mean private Cloud can be a very good buy in certain circumstances.  But you just need to do, and Government needs to do this assessment I think of the total cost of ownership of something to realise the benefits that you might get.

Elasticity is an important part of Cloud Computing.  That’s not to say that you might want to use some aspects of outsourced provisions that don’t have these characteristics, but the notion of Cloud is the notion of being able to run up to what you want, expand, perhaps not quite to infinity, but certainly a very long way, and then wind it down again so you’re not paying for the things that you’re not using.  And because it’s that defined point, this notion of a measured service, one that I get because I pay for what I’m using rather than securing the service in the first place is a very important arrangement.  Now I think in public Cloud that’s pretty easy to understand.  In private Cloud if people are doing it properly than that’s also fine, but sometimes you run across vendors who are saying well there’s actually a premium to get into the service, you need to do something to set it up to start with, so you never not actually paying nothing, but you can use more, but there’s a base payment.  Now that’s not to say that’s a bad thing, and it might be cost effective to do it, but it isn’t pure Cloud and I think it’s important to draw those distinctions so you understand.

Now the next thing I think that it’s important to understand about Cloud Computing is it’s not a religious experience.  It’s about procurement.  It’s a different form of procurement, buying different technologies than we’ve been buying previously, or perhaps combined in a different way than they’ve been combined previously.  But it doesn’t require you to abandon everything you’ve done before and move wholesale to the Cloud.

There are some interesting things to think about in the context of buying Cloud Computing.  One of these is the notion of using storage in the Cloud.  Now what do you know about Government data? And basically we never throw anything away.  I think the minimum time range for keeping something is about seven years, that’s our common experience in tax records and those sorts of things.  If my memory serves me correctly military personnel records have to be kept for 130 years.  Is Amazon going to be around in 130 years?  I don’t know.  I hope so, they’ve got all my books and I want the grandchildren to read them.  Perish the thought.  There’s this sort of challenge of what you’re going to do and what are the right things to put in the Cloud.  The notion, I’ve heard very interesting discussion by some banks of keeping their disaster recovery in their own data centre and are running their operations in the Cloud, so if something goes wrong they can get back and continue to work, rather than depend on something else.  And I think that’s an interesting way of doing things.

The other notion is the idea that you don’t want to be… and I’m going to use Amazon as an example, you don’t want to be in just one region, because what we’ve seen in the Amazon ECC failures have been one region going down, but if you’ve been across more than one region it hasn’t normally been a problem.  And this is no bad thing because we understand that for critical data, critical data centres, the notion active arrangements is a very strong one, or the notion of disaster recovery somewhere else rather than where your main stuff is.  They’re normal arrangements.  And as a consequence when you’re looking at what you might put into the Cloud and what that guidance talks to you about is looking and examining these things to make sure that in moving to the Cloud you haven’t become some sort of zealot, but rather you’re remembering to balance the things that need to be balanced.

Our recently published implementation guide for Cloud Computing, and this was the most recent in this series of things that we’ve published, uses this diagram to talk about the opportunities and risks that you need to consider in doing Cloud Computing work.  And I think it’s a very good example of the fact that this is a normal procurement decision that you might be making.  Now I’m not in favour, and this is my personal opinion, I’m not in favour of a Cloud first policy, or indeed any other sort of first policy except under the Commonwealth procurement rules, the notion of value first, because that’s what the Commonwealth procurement rules say.  Pursue value for money, that’s the standard by which you decide what to do.

Now that sometimes means that you won’t be in the Cloud and it sometimes means that you will be in the Cloud, but I think some of the notions of Cloud first policies are probably not as strong as they could be, but I think we do need to encourage people to look at Cloud Computing as a way of doing things.  And I think that’s an important challenge, to change their minds and if you want an example of how we’ve done that, we did it with the Open Source Software Policy where we said in the Open Source Software Policy that what agencies had to do is show that they’d considered open source solutions, when vendors presented proposals to show whether they’d considered open source solutions and what they’d done, not to say that a particular thing is useful every time you do something, because clearly it’s not, but rather to make sure that people have considered the benefits of these things.

Now that looks a bit, I think, like a risk-averse policy if you don’t understand it perfectly well, but I think once you get into it it’s about being aware of the risks and making decisions that balance risk and look at the rewards that you get as a consequence.  And this is an important factor when we take into account what we’ve done in the Data Centre as a service multi-use list.

Now I’m going to spend a few minutes talking about the various details of this arrangement and then we’ll be in a position to sort of wrap up and then look at questions.  One of the challenges that we, after we finished the sort of the way that we’ve historically done procurement in the Data Centre strategy, so the notion of panels for facilities and migration services and targets to be measured and things like that, was what do we do to take into account Cloud Computing?  What can we do that will assist agencies in order to take up Cloud things quickly and easily?  And of course the first thing that you turn your mind to is the notion of services that aren’t covered procurements, so they’re below the tender ceiling of $80,000.  And when we started to look at the Data Centre as a service work, we said well let’s limit risk, we’ll cap it at $80,000.  We’ll cap the duration of one of these contracts at 12 months.  We’ll say that there’ll be compulsory arbitration instead of the first step being going to legal arrangements.  In this multi-use list, unlike most multi-use lists, the vendors who join it will be required to sign a contract committing to all these things.

Now the multi-use list can still be, and has been, used as multi-use lists have been used historically previously, as a method for downsizing for a select tender.  And you can use the list for that, and indeed one agency has already.  But to buy off the official order arrangement, the speeded up procurement arrangements in the Data Centre as a service arrangement, what you’re limited to the range this can do.  Now below $80,000, it’s the Chief Executive Officer’s Operating Instructions, CEIs for agencies that determine do they need three quotes, do they need ten quotes, do they need them in writing, do they need that.  But that’s for the agency to decide.

What Data Centre as a service provides is a way of getting those quotes relatively easily.  We didn’t set an RFT arrangement in the Data Centre as a service.  Instead of going and saying we want these qualities and this security and those service levels, we said here’s the range of service levels that you might have, describe your services by picking one of those that matches.  Can you set up your arrangement in less than an hour?  Does it take a week?  Are you providing 24/7, 365 support, or are you providing 8-5, 44 support, or something like that.  No compulsion to do anything else, but just tell us what the service is like.  Is your service in Australia?  Is it certified in a security sense?  What are the help desk arrangements that you have?  All those sort of characteristics are outlined, and the people who contributed to the work chose what it is that they wanted to do, and how they would offer their services.

Now because we recognise that different levels of service levels drive different cost models, then you offer your services at these different models, so you can charge more for the more secure one, more for the one with higher service levels, that’s what you can do.  So an agency can say well I want this level of service, I want these capabilities and then broadly break it into infrastructure as a service, platform as a service, software as a service and some others.  I want this sort of categories, you heard me talk about CRM, but there’s web hosting, a whole bunch of categories as well.  And it’s all in a publicly available, the prices aren’t, but a publicly available, drop down spread sheet.  Government agencies can see the price, but the public can’t.  But you can choose exactly the service that you want, and sometimes you’ll get a list of providers who do that.  Sometimes you’ll only get a couple, sometimes you might only get one, but you can then go and get that service, they’ve already signed the contract arrangements, so you can go and buy that service and have it set up very quickly.

All of this is designed to reduce the risk.  Over time we anticipate that agencies will report, again a very simple reporting mechanism, on how they have used that service, whether it’s been successful or not, and we’ll record those sorts of things in the database as well, so people will be able to say well this is a pretty cheap service and I like the look of it, but I notice that ten out of eleven agencies that have used didn’t think it met the requirement, maybe I ought to go a bit higher in price level, or something like that.

So it’s a simple, a service catalogue arrangement, we had 35 providers offer 945 services in the first tranche of activities.  Another 25 or so providers, around that, have offered services in the second tranche that we’re just evaluating now, and they’re offering another 500 odd services or something of that order of magnitude, so clearly there’s a bit of demand.  We released it in the middle of October, we’ve had three contracts since then, and one agency as I said that used it as the basis for a tender.  Another two of those happened to have been projects of mine, but in another situation we needed a test in Dev environment, so we went through the process.  It was the first time this particular branch had used the panel, so I had to introduce my two branches to each other and get them to work together and do that stuff.  That took about a week to get the quote out.  We signed… they rang up to make sure that the vendor knew what was going on, when will we have the test in Dev environment available?  The next morning, before we could get the programmers, the developers on board.  That’s as fast as we’re looking at turnarounds.  And they’re things that haven’t previously been able to be done across agencies.

Now these aren’t environments that we’re going to keep for 100 years, as I said.  They’re not all multi-level security and things like that, but it shows a way of getting out there and making use of this new way of buying IT that hasn’t previously been put together all at once.

The obvious question is what will we do now?  First of all we’re going to see if this works.  We’re going to see over the next two years if agencies do take it up and their interest in pursuing it.  We’ll then look at whether or not we need to do some form of providing the resources, at the moment I’m paying for it out of Data Centre Strategy money, but eventually that will run out and I’ll need to look at how we run things, but I’m hoping that it might be a bit more automated by then, so that would be useful.

There’s this interesting $80,000 limit.  There’s clearly a lot of computing that agencies would like to do that’s at a higher level than that, and of longer duration, and we think the answer to that is probably going to be around taking a second tier of provider through a tighter contractual arrangement that looks at other things, so they’ve got a different head agreement that looks at other arrangements that sit above that $80,000 ceiling, perhaps we might have to do some more work in an evaluation sense to look at those vendors, the more sort of historical things you do in normal tender evaluation to get them into that second tier and maybe we’ll be able to look at something like that.  And then you’d be able to extend the options for contracts more than one year, renew them and those sorts of things.  That’s what we’re thinking at the moment about how we should do it.

Now I would say in sort of our defence that a lot of that stuff that you read in the IT press about AGIMO’s slow attitude to taking up Cloud Computing and our risk averseness and all those sorts of things I obviously think is largely rubbish, written by people who aren’t responsible for what they’re delivering.  It always amuses me to read articles from advisers and things like that that say AGIMO should do this, and AGIMO should do that and you read the disclaimer at the bottom of their thing and it says of course this isn’t advice, don’t think it’s advice, if you act on it you can’t sue us, things like that.  Well there’s a bit more of a problem if your job rests on these things, I think it tends to focus the mind.

I think we’ve got the balance about right, but I’m interested in hearing about what we can do to improve it, what we can do to change it, and obviously we’re going to be talking to industry and agencies just as we have over the last couple of years about this, to make sure we can drive things better into the future.

Questions?

Pete: Yeah John.

John: Pete.

Pete: How will you measure success in about two years’ time?

John:  We’ll look at the number of people who’ve used the panel, that will be, I think, the first one.  We’ll want to look at the results they’ve had.  We’ll want to ask the vendors if they’ve been getting enough business from their point of view about those things.  Some of you would be aware that recently we closed two multi-use lists that had about 800 vendors on them.  We asked the 800 vendors what they thought, 218 replied, I think something like 90% hadn’t been getting any work from the panel.  Agencies hadn’t been… I think two agencies out of 30 something said they’d ever used the multi-use lists, and we closed them down because I don’t want to waste resources doing that.  I’ve said about this if no one uses it and it doesn’t turn out to be a good idea, it’s not a mandatory arrangement, we’ll just turn it off.  But my view is that unless you take some risks in what we’re doing, unless you look at some new approaches and try and do things, you’re just going to get the same results.  And I think it’s sort of incumbent upon us to be a bit innovative, particularly in IT when the world’s changing, to see what we can do.

Question from the floor: John, I applaud the approach, I think it’s fantastic.  My business, local government, state government, boy, I wish there was some leadership like this.  That’s a rabble in those areas.  I guess we still in Federal Government hear a lot of questions about we’re not quite sure what we want to do, how are you going about educating the market in this process?

John:  It’s interesting, I heard to my horror the other day that one vendor went to speak to four agencies who’d never heard of Data Centre as a service, including one of the agencies that had replied to some of our work.  The challenge for us is getting out there and telling people about it.  So we do a range of things, there’s the delightfully named SPORG, the Senior Procurement Officers’ Reference Group which I like because it sounds like it’s in (inaudible – 40.39), that we talk to every three months.  We talk to another activity, a larger sort of 200-300 people in procurement forums.  We put procurement newsletters out.  You can read about us on the AGIMO blog.  There’s a whole range of doing those things.  Do we get as far… how do we get as far as convincing people and persuading them to do that, well it takes a while to get that.  People are always interested in new approaches.

As I say to Data Centre providers I regularly have discussions with Data Centre vendors that aren’t on our panels who come to me with enormous great charts of Government demand and say “Look at all this demand, you should add us to the panel,” and I sort of say, like you understand that some of these people are just kicking the tyres don’t you, and they’re not giving you the whole story and sometimes you miss that.  I think what we have to do, what we have to keep doing is telling agencies what we’re doing, publicising it in things like this and other ways, I also think that the vendors have now got some incentive to go to agencies and say well did you know I could provide this service for you through this very simple arrangement, and I think that might be useful as well.

Question from the floor: You mentioned there’s a bit of a cycle in terms of who gets engaged in these discussions, how long before you think the lawyers will actually get into the discussion and add some value?  And where do you see they can add the most value?

Question from the floor: They could buy the drinks.

John:  No, we do a lot of our work, and what we’ve done with the multi-use list contract, we put that out on our web for consultation, we spoke to the Finance inside lawyers, they did a lot of good work for us in taking into account the recently written under $80,000 contract and modifying it for this.  I had various healthy and frank engagements with the lawyers about the notion to make sure they’re instructed properly and then the business gets to decide the risks, so we had a very interesting discussion on compulsory arbitration and the words we would use to put that in the contract arrangements.  They provided their advice and were very helpful about it, and we got that built into the thing, so there’s arrangements to do that.

Clearly, and we’ve got a whole advisory guide on the legal issues to do with Cloud Computing and it’s worthwhile looking at those, and making sure if your agencies are doing something, that if they’re not doing it through one of the list or panel arrangements, that they’re aware of that guide and look at those sorts of things.  One of the interesting challenges about public Cloud is the notion that an agency with 300 people can negotiate with Amazon or Google.  Good luck with that.  Those public Cloud providers aren’t going to change their terms and conditions for those arrangements, and you’ll notice that we’ve got some large companies in the first tranche of providers on Data Centre as a service, but not some of those big public Cloud providers.  I’m confident that over time we’ll get more of those involved as they look at ways to get into this sort of business, particularly as it looks like Government is taking it up strongly.  But like any procurement, you need a sort of appropriate dose of legal advice.  There’s a question up the back there somewhere.

Question from the floor: Yeah, we’re just wondering what the timeframe is for the above the $80,000 cap, if you can comment about how long (inaudible – 44.17).

John:  Yeah.  It’s an interesting challenge.  I think, as I said, we put this out… turned it on in mid-October.  It’s taken us quite some time to get there, to get things right.  I think I’m inclined to wait and see how it goes for a while first.  I don’t think I’d do anything in less than 12 months.  One of the experiences we had under the Co-ordinated Procurement Arrangement is my division conducted 11 whole of government procurements simultaneously from the period 2009 through about the middle of 2011.  Some of my staff don’t think that’s a particularly good idea and don’t want me to do it again and I sort of understand that, so I think we’ve got a range of other procurements that we have to turn over in the next 12 to 18 months as their initial three year periods expire.  So I’m thinking a year to two years we’ll sort of look at that.

Chairman: John, I think I can be fairly confident and say that most project managers in this room have experienced either receiving a project brief or being handed a project where the deadline has been set.  You know usually a commitment’s been made to the Minister that the grant management system is going to happen on the 30th of June, or whatever, and I think what you’ve outlined today is, and I like your term, is a procurement option, a procurement option for situations like that, where it may not suit all situations, but you’ve outlined the technologies, the capabilities that we should be looking for, and you’ve put in place annals to help project managers put together those solutions and so I think it’s been a very useful hour, and I appreciate you…

 

Comments on this blog are now closed. Please let us know if you would like to discuss this post or have any general comments.

Last updated: 05 August 2016