1
AGIMO – Data Centre as a Service – Industry Briefing 2 May 2012
Introduction by John Sheridan
Welcome ladies and gentlemen to our presentation today on data centre as a service. This is a range of things that we often do in terms of industry briefs but it’s the first time that we’ve done one that’s been webcast live and in which we’ll be taking questions from not only the audience but our audience online. For those of you listening online, the hash tag we’re going to use for this purpose is hash tag AUSGOVIT, and we’ll be welcoming your questions probably about half way through the hour we should be ready to take questions then. There’s also the possibility of emailing questions and delivering them in other ways through the blog. We’ll be answering all the questions that we receive this way on our blog if we don’t get to them all today. We’ll finish today at 5o’clock but there is the opportunity to ask more questions and get them answered. We think this is important because we want to make sure we consult fully on this new way of doing business for government.
There’s been much in the online press recently about AGIMO’s work in doing strategy of various sorts. Two of the strategies that affect the work that we’ve been doing are those in the data centre area and of course in the cloud area.
Now these two strategies were issued bias some considerable time apart and it would be reasonable to think where’s the direct connection between the two strategies. First of all let’s just briefly discuss what the data centre strategy is about.
The data centre strategy is about avoiding costs over the next 10 to 15years. So Sir Peter Gershon in his review done in 2008 worked out that there was a possibility of avoiding about a billion dollars worth of costs in data centres for government over the next 10 to 15years. That doesn’t mean we won’t spend money on data centres. At the time we were spending about $870M a year on average for our data centre business. And we do anticipate that those levels of spending are going to continue into the future. But what we see is a means of making sure we’re not spending anything that we don’t need to.
Now for those of you who don’t deal closely with the budget ramifications of government, I’ll just briefly explain that costs avoided are money that you don’t approach government for. They’re not savings in the sense that they’re reduced from an agency’s budget, or taken from the agency in anyway, it’s the ability for the agency to do something without having to go to government for extra money. And that’s what this strategy aims at, the data centre strategy is aimed at.
Now since then of course the rise of the cloud arrangements, mean that it’s quite possible for people to pursue data centre savings or data centre cost avoidance in a way that doesn’t require them just to get the best possible data centre facilities, because that was the focus of most of the data centre strategy.
What the cloud enables us to do is to take advantage of other savings opportunities that don’t require us to build or lease or otherwise procure facilities for the long term. These are very much complementary strategies. They don’t run against each other. We don’t actually see much competition directly between the them. What we see is these two arrangements allowing us to develop a better arrangement and a more efficient spending of government money over the next 10 to 15years as we look at how we bring IT into the usage in the mid 21st Century.
Now we do anticipate that we’ll need to get more, or continue to get the level of cooperation across agencies that we’ve seen demonstrated in a range of coordinated procurement activities. Some of you would know that in the data centre facilities area, the leases we’ve signed up have been a consortia of agencies, not just a single agency. So we get the benefits of the aggregating of government demand and the economies of scale that that allows to get better prices and better arrangements for our work. Now we expect that the data centre as a service project will do the same sorts of things.
Let’s turn to it now. You would probably know that there are 111FMA Act agencies. Of those 111, more than 50 spent less than $2M a year on IT. Now the risk of repeating myself for some of you who heard me speak yesterday, you need to understand what that means.
There are probably four people in that IT shop. One looks after the email message arrangements, one looks after applications, one looks after procurement and one is notionally the CIO. And that’s all there is. And they spend a lot of their time doing things that are common across agencies, not things that are specific about their agency.
And what we want to do in the data centre as a service arrangement is continue to contribute to the work that we’ve been doing to reduce the amount of common repetitive work that agencies need to do, particularly in the area of procurement. You would know that going to market to procure things, particularly things that are covered procurements more than $80,000 a year, is a long and often involved process. I’m not saying that it’s not effective. It get’s value for money but it gets it in a way that puts a lot of effort onto smaller agencies because they each have to go around doing something about this.
Now our work in procurement policy and ICT services panels, has started to address this by reducing the number of panels across the FMA Act agencies, ensuring that each portfolio has a maximum of three panels, and that each of those panels have piggy back clauses that allow them to be used to get the best benefit across a range of agencies. Eventually we anticipate that not all agencies will want to have their own panels.
Now in the data centre as a service area what we’re trying to do is also address the more than 70% of ICT services contracts that are less than $80,000. A lot of the work that one sees can be done in the cloud now, allows us to look at smaller levels of contracts. Most of the time these arrangements, less than $80,000, are not covered procurements in the context of the “Commonwealth Procurement Guidelines”, but are instead subject to agency chief executive instructions. The agency might require three quotes, three written quotes to be done for a particular service.
Now at the moment, as I’m sure many of you would know, if you’re not known to the agency involved, the chances of getting a quote, asked to be asked to quote, aren’t particularly high. Particularly in areas where there is new technology involved. Somehow you’ve got to get to the attention of the people who are making these RFQ decisions if you like, so you can be asked in the first place. And if you are asked, you need to be able to prove your reputation, you need to be able to demonstrate often from scratch that you can do the things that are asked. You don’t necessarily have the advantages, particularly if you’re a new firm, that some more established firms have. What we’re trying to do in the data centre as a service work is change that arrangement to increase competition, to address these areas of commonality, to provide a better approach to the market and a better market based response arrangement so that we can encourage new growth, we can increase competition and we can reduce the transaction costs for both agencies and vendors in providing these services.
Now we started doing this work sometime ago in 2011, and as you would expect conducted a range of activities, firstly with agencies to determine what it was that they wanted. We didn’t set out to put in a mandated arrangement to force people onto a particular panel. Because we know that in the absence of firm data about demand, and particularly when there’s a plurality of suppliers, it’s very difficult to mandate what it is that people are going to do and ensure that the people on the panel, the vendors on the panel get the work that justifies them responding to the tender arrangement in the first place.
There’s also a lot of work to do when there’s a lot of possible vendors, in getting a shortlisted panel arrangement because of the need to assess in great detail the offers made by those vendors. So it was clear to us that we needed to do something new in this area.
We also spoke to industry on more than one occasion. We conducted an industry briefing, we invited a range of companies to present to a sort of cross agency panel of our procurement and other experts in this area, and we learnt a lot about what was being offered in the market. And you can see some of the points there on the slide.
It was interesting to see that the challenges of dealing with government, of dealing with government procurement rules did still affect a range of vendors. And the idea that we would somehow reduce that burden was an important part in deciding how we’d go forward.
So we came up with the current approach that many of you I trust would have seen discussed on the blog and in other ways since we put it up. I’ll just touch for a moment on the notion of consulting on the blog and what it means to us. Some of you would know that this is one of the posts that data centre as a service post, is one that’s got a lot of comments from a range of people. And we very much welcome those comments so that we can have the discussion in a way that I believe isn’t particularly confronting, that can explore points particularly in this early stage and help develop our thoughts. Before we had the opportunity to use the blog for these arrangements we were stuck with a problem of probity. The probity requirements in a nutshell essentially said you can tell everybody everything or nobody anything. And since we had no means to tell everybody everything, we ended up telling nobody anything, and we gave industry briefs that frankly were, I would suggest, not particularly useful. We often started the brief by saying nothing you hear has any bearing on what’s actually in the paperwork, and one got the feeling that people were only coming to see their old friends to those industry briefs.
What we’ve got instead is the possibility of getting comment from industry and indeed from other stakeholders in a way that allows you to declare who you work for, say what company it is, talk about your company’s points, or indeed do so anonymously if you would prefer, make the points.
We’ve offered the opportunity, and indeed you will have seen that we’ve done that, to allow people to email in, to get comic questions asked and sort of we’ll self ask the questions as a consequence of those emails, and provide commentary on the questions that have been asked for us. We think this is an important part of using Gov 2 tools to engage with out stakeholders. And I’d like to thank the people who have done that to date on the blog, because it’s been very useful for us. Some people in particular have gone to a great extent to identify problems with individual clauses that they saw, and make suggestions about adjusting them, and that’s been very useful in helping me form my own ideas about what might work in this regard.
Now let me just touch on a couple of important things about the approach. One of the questions we have had has been, why did you limit it to $80,000? Well I think we know that in the cloud setup at the moment you can actually buy a lot of computing for $80,000. But the real reason is to reduce the risk, and thus reduce the amount of effort people have to go, the transactions costs of doing business on this particularly multi-use list.
Another question is, why would you have a multi-use list rather than a panel? And the answer here is also that it reduces the amount of work that has to be done to get onto the arrangement in the first place. And I’m going to touch on some more of that details shortly.
Now it still falls within the overall Commonwealth Procurement Guidelines or Rules as there be from the 1st of July I think, the need to get value for money, that delegates must get value for money remains. But what doesn’t occur is we haven’t gone through this covered tender process in order to make sure that everybody who gets on a panel was in fact prequalified. What we’ve done instead is arrange what I would see as something of a lower hurdle. Now this hurdle will require us to check a few obvious things. We’ll need to know that you’re a real company if you apply, and we’ll probably use a service like Dunn and Bradstreet in order to do that. We’ll also need to make sure that 10 applications aren’t actually from the same master company and they’re creating sort of false competition between them. And again those company checks will help us do that.
We’ll need to make sure that you’ve got some references, or something that can be used to establish your qualifications to do work in this particular area. We think at the moment that sort of two references will be the way that we’ll do that, but I’m open to ideas about how better we could do that. And particularly regarding ideas about how we would get new entrance into the market here who might not have those references.
There’s an interesting question, by the way, about whether, given the nature of government procurement, the first job someone gets in the data centre as a service area should be working for government, but that’s something that we’ll be able to take into account.
The other arrangement, and I think it’s the important one that I want to get across here, is that we’re not just doing the normal multi-use listings, we’re also putting in place or intend to put in place a head agreement or a deed that vendors will sign up to as a condition of being on the panel. Now what that’s designed to do is to take out of the, if you like, the normal contract negotiations that are required in order to get a service, all the discussion about the things that everybody has to do anyway. Equal opportunity arrangements, workplace health and safety arrangements, those sorts of things, and get them signed up.
It’s our intention that, once following the consultation we’ve had today and in previous weeks, once we get that head agreement right, it won’t be negotiated by companies with us. Companies will sign the head agreement or they won’t. And if they don’t sign the head agreement then they won’t be able to offer services under this multi-use list arrangement. They’ll still be able to offer services in other areas and go through normal tendering procedures in order to do that and negotiate away to their heart’s contents with agencies as regards that, but this particular mechanism won’t be open to people who don’t sign the head agreement.
Again that’s why we’re keeping it under $80,000, because the amount at risk for the commonwealth isn’t particularly high. Obviously we also know that for $80,000 the commonwealth will most often be paying in arrears as regards this. That means that the amount of money at risk for the commonwealth is relatively limited.
Now I’m sure all of you know the cost of lawyers in order to get things through the courts. And we know that frankly, and I’m sure this isn’t a secret, that we’re unlikely to want to drag someone into court to recover a payment of less than $80,000. It will take us longer to do that than will be worthwhile. What we want instead is a range of ways to put things in place so that we can address this in a better way. We think that to do this in the first instance, we might have a compulsory arbitration clause that says the parties agree to use arbitration in the first instance, sorry not in the first instance, in the penultimate instance if you like, if you can’t resolve things other ways to use arbitration, independent arbitration to resolve matters. And the reason that we want to do that is to, so we all agree that we’re not going to go to court about these things.
Now I’m sure you know that we can’t actually mandate not going to court in those circumstances, but my advisors tell me that if we set this arrangement up, the people who run courts will take a dim view of people who didn’t go through the arbitration arrangements first. And we think that will help address disputes that always arise in contracts over time.
Now the list itself, we’re planning to put it in place for two years. I’ve got funding to cover the first year of this at this stage, and I’ve got a means of exploring funding to cover the second year. But what I don’t want to do is put in place an enduring arrangement that no one uses. If it turns out that this doesn’t work, we’re prepared to say well okay it hasn’t been an enormous burden for vendors to get onto it. It’s not like they had to go through an enormous RFT process. We haven’t charged them an enormous fee in order to that, and I’ll just come back to that in a second. We haven’t charged them an enormous fee, so if it doesn’t work and people aren’t getting business, then we’ll be able to, if you like, wrap the thing up and say well we tried to do something different, it didn’t work, we’re going back to other ways of doing things. Now I hope that that doesn’t happen. I think this is a strong possibility of addressing these sorts of ideas quite well, but there’s always the possibility.