Spreadsheet – desktop or server?

I currently see spreadsheets as a desktop based application. But the more I think about it the more things are quietly moving to the server. And have been for a while.

Pretty much all the spreadsheets I have written in the past few years have pulled some data from a server somewhere. Apart from one app I remember where all the ‘data’ was made up by someone else in another spreadsheet. (There was no real data anywhere).

The add-in type stuff I have written has gradually moved to a more centralised deployment approach, generally network file based rather than pure ‘server’, but getting there.

I was thinking that we need the processing power of a decent desktop, I have some horrible spreadsheets that take minutes to calc, and VBA that takes hours to run. I was thinking that couldn’t be done over a network because there is too much data there would be too much latency. I am used to pitiful client networks where it can take 5-10 minutes to open and save large models (which means that Autosave ‘virus’ basically locks you out of the app almost all the time – well it would if I didn’t always disable it).

When I thought about it though, if the spreadsheet was living on the server, the network traffic would/could be small – just transmitting commands. This is just how Citrix works for example.

So we have data gradually tending to come from a server, we have command based add-in logic from a server, if we put the spreadsheet logic on a server too, it can do all the hard work in the cloud before it lands on our desktop. This would stop us being locked out of doing shopping lists just because the billion dollar deal evaluation model is calcing.

Harlan made a good point that spreadsheets aren’t especially the best way to implement black box business logic. But I can see the appeal if its something the business users can work with themselves. And with it being server based it can be migrated over time to some other technology that may be more appropriate.

Personally I love the autonomy of having everything installed on my pc so I can work when/where I want. But I also love the simple update deployment process of server based apps. I can imagine things grinding to a halt at period end times though as everyone tries to calc their spreadsheets monsters at the same time on the same server. But maybe if duplications were removed a lot less work would need doing?

What do you think, are spreadsheets locked to the desktop, or can you imagine that they will be served up remotely eventually (/soon?)?

I am not suggesting any particular technology is going to solve this, it could be Excel Services, could be Google spreadsheets, it could be a server based xll host. All have pros and cons, although I suspect those that are easiest based on what people already have will have significant advantage. What do you think will be the key factors that drive adoption? or why don’t you think it will work?

cheers

Simon

Advertisements

16 Responses to “Spreadsheet – desktop or server?”

  1. MacroMan Says:

    So I guess Excel Services is the future?

  2. Ken Puls Says:

    Interesting…

    I love Citrix. I have two servers in our farm at work, and it’s the best damn platform since sliced bread, IMHO. I can deploy an app enterprise-wide in minutes, rather than installing on desktops everywhere. We should make very sure, though, that we don’t confuse Citrix with eliminating desktops though. The apps are still there, and run as if they were on an individual’s PC.

    One REALLY cool thing about Citrix is the secure gateway. We just recently set this up, and can now log in from anywhere in the world. Using on Internet Explorer, I can access a full real-time desktop, just as I would get if I sit in my office. Same file access, same rich Excel application and no noticeable difference in speed. Absolutely awesome. Technically, these are both locked to the desktop AND served up remotely.

    It’s actually the secure gateway that made me start to loath web applications. Using current web setups to deliver applications that require any kind of real time interaction just doesn’t cut it any more. Our bank’s website is one that really gets me going… they use that horribly slow, miserable excuse for a language called Java. Every page takes the better part of 5-10 seconds to reload, and it is designed to be as inefficient as possible, I think. While the latter issue is a problem that could easily be fixed, the other requires a shift in technology somewhere. Who knows, maybe they could put it on Sharepoint and it would be better.

    Personally, I see server based applications being used in the future to push out canned reports with a very limited numbers of options. The Excel “Rich Client” will be used by those with enough skill to develop these kind of reports and do the ad-hoc data entry. To my understanding, this is what Sharepoint and Excel services is supossed to offer, isn’t it? Canned reports with a few options, and that’s it?

    As far as adoption, I figure it comes down to a cost/benefit thing more than anything else. For us, Excel services with Sharepoint is not really a compelling thing. We have standard, canned reports, but we also have full Excel installs available to everyone, which they would still need to do their job. The only reason I can see to publish using Sharepoint currently is to lock data down, but it requires a Sharepoint server to do it. For the sake of my user base, that’s a heavy cost to pay when I can just PDF my reports for nothing and send them out.

  3. dermot Says:

    I used Citrix for some time last year. I found the internet latency to be a problem (it was like working underwater – you had to slow down), also it seemed to process keystrokes before mouse clicks, so I’d click and type and it would go in the wrong place.

    I can clearly see the logic of centralised control, but the spreadsheets I work with don’t stand still long enough to become immutable standards. And as long as spreadsheets continue to be the scratchpad/sandbox of business, centralised control just doesn’t work for them. If you’re talking about “systems”, that’s different. Excel wasn’t built for that, imho.

  4. gobansaor Says:

    In most organisations “server” implies central IT control which implies intra departmental budgets/cost sharing. Unless server based solutions offer a significant benefit to offset this added “cost of doing business” (i.e the endless meetings, ROI calculations, extra IT staff etc.) , then desktop based spreadsheets will continue to be the preferred choice.

    Tom

  5. Marcus Says:

    While I like the idea of a server based calculation engine (not necessarily Excel Services) I think Tom’s on to something here.

    In a recent project I was involved in migrating a spreadsheet based modelling tool to a production environment (Informatica, SQL Server and VB6 DLL’s). The business’ main grievance was the loss of control they had now that their baby was on a server.

    “love the autonomy of having everything installed on my pc”
    Most business power users also love this autonomy. And if they also happen to be discission makers, that’s probably where they’ll stay.

    Shahar (if you’re there) – have you noticed any trends?

    Regards – Marcus

  6. Simon Says:

    Macroman – no it will be a cultural challenge, yes it may be better than the current situation, no it is not ideal! ie I don’t know!
    Ken so is the question what benefit does Excel Services offer over Citrix? interesting thought
    Tom/Marcus I agree 100% on the control thing, my sense is that the more switched on finance departments are getting server access, for eg Essbase departmental SQL servers etc. Are you seeing that?
    I guess I generally work in the finance systems dept so could have a skewed view.

  7. Marcus Says:

    I don’t know what the trend is in the UK but here I’ve found the more ‘switched on’ departments have done two things:

    > put their hand in the pocket and paid for a dedicated server for MIS

    > jumped on MSAS (over Essbase). In a prior project we migrated an Essbase data mart to MSAS because, well it’s free. I believe the business figure that the migration would pay for itself after 12 to 18 months (saving on Essbase’s server and client licencing costs).

    Cheers – Marcus

  8. Simon Says:

    Dermot good point of the constant change, I can imagine scenarios where a server spreadsheet could be good and bad, and even unknown.
    I had a slight delay on Citrix too amde it hard to type, and it never loaded my Excel so it moved down on enter, didn’t have my add-ins installed etc. OK for some probably, but not power users I suspect, like so many of these things.

  9. Marcus Says:

    Simon, you are a night owl. It must be 1:45 AM over there.

  10. Simon Says:

    Marcus – trends – as ever – http://www.jobserve.com/
    (no Excel services mentioned)
    AS v Essbase, I’d like to imagine that people reviewed the features asked around and chose on which had the strengths they needed. But as you say Free is pretty hard to beat.
    I think it will be interesting to see if there are more or less of these migrations now Oracle are at the helm. I know Hyperion had just made some unwelcome licensing changes, which may have put AS on a few peoples radar.
    So do you think the ‘server control’ factor may be easing off as a barrier to some of this stuff?

  11. Simon Says:

    spot on Marcus! Are you on your dinner break?

  12. Marcus Says:

    Ironically many expenditure decisions are based upon how things get shuffled around on the balance sheet or P&L. If you going to spend the money, arguing whether it should OpEx or CapEx is just shullfing desk chairs around on the Titanic. Here the business was happier putting the money into operational staff costs rather than licencing.

    ’server control’ factor may be easing off
    Difficult to say. Part of the equatin is flexibility. Many exec’s take work home or to a client with their laptops limiting the attraction of the server side approach. One nice feature Excel had (I think it’s no longer there) is the capacity to create local cubes. This made it easy to create a solution which could switch between a server based or local data MSAS cube.

    Cheers – Marcus

  13. Marcus Says:

    “trends – as ever – http://www.jobserve.com/

    Meanwhile though the London VBA market appears to be quite healthy.

    http://www.jobserve.com.au/JobListing.aspx?shid=CB1D958B63DE75

  14. Marcus Says:

    No – it lunch time in about an hour (11:10 AM Friday morning)

  15. Simon Says:

    get back to work then!
    Yes plenty of Excel VBA work in London

  16. Harlan Grove Says:

    Re decision-making power users: if they make good decisions, they gravitate towards making more and more decisions, less and less development. Power users who don’t make decisions gravitate towards obscurity.

    Spreadsheets aren’t going to survive the retirement of those who learned how to use them when they were new(ish), in the 1980s. I haven’t met any under 30 developers with any spreadsheet competence, but my perspective may be unusual. Besides, back in the day I had 123, DEBUG.COM, BASICA, Turbo Pascal and batch files. And I used Turbo Pascal for my first significant work. Today there are many, many more alternatives, and with regard to new kids, I’d trust universities, colleges and technical schools to do a better job teaching scripting languages and databases than spreadsheets and anything or nothing.

    Not all that long from now, the decision makers will be used to browser-based decision support systems. Maybe a few of them will use spreadsheets for the odd ad hoc analysis, but it won’t be the tool they reach for first.

    And [re]centralization is inevitable. For me, 90% of the non-development part of my job I already do through Citrix. Why? Because it ALREADY DOES cost less than trying to distribute and install the same apps on local machines. Developers and the few (fewer than half) employees given laptops may prefer freedom and flexibility, but the majority of users with standard, not easily transported desktop machines could be centralized without much fuss. And in a few more years WiFi and VPNs are going to be so common that anyone with a wireless network adapter isn’t going to feel much loss of freedom or flexibility from centralization either.

Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


%d bloggers like this: