> The obvious advantages and disadvantages of SaaS I would percept are: > Advantages: > 1) No onsite administration – lowers cost of ownership
This is not an on-premise versus SaaS issue. Many people work remotely on "on-premise" systems. Nothing good VPN software can't handle. > 2) You are almost always up to date on versions etc. Again, this is not an on-premise versus SaaS issue, but one of the degree to which you customize the core product. The more customizations the longer the testing cycle to see what breaks and what is replaced by new functionality. SaaS CAN be customized if the vendor allows you to (which is true in all of the cases I can think of, which are not that many). > 3) You do not risk downtime when a system is upgraded, or during > system maintenance, or bug fixes. Ummm. No. Many of the ITSMaaS products have the same risk, but is often reduced simply because the level of customization is generally lower. > The vendor usually has a faster planned route to rollback. That may be so, as they go through the pain multiple times (because they have multiple customers), whereas you tend to only do it once. I think they have a better plan because their pain is higher if it fails. > Disadvantages: > 1) No onsite administration – reduces flexibility in some areas of > customization. You'll have to explain that one. The tool itself is no more or no less customizable because it runs on-premise. Maybe you are referring to it being easier for developers to collaborate with customers on customizations? If so, being close to the customer is probably a "bad thing" as it leads to excessive "tweaking", which in the end makes the upgrades so painful. > 2) Your data resides off premise so it poses some kind of security risk I hear that one a lot, but I do not buy it. Possessing data poses risk, period. People who specialize in the security of data centers tend to have a lot more knowledge of how to do it than your average corporate IT security guy, who is probably wearing a hat and a half. Terramark, for example, is Federal Information Security Managements Act (FISMA) certified, which is why Federal government IT is starting to seriously consider SaaS solutions that house their data with them. I can think of a few Federal operations that would not pass a FISMA certification audit. > 3) You are vendor/manufacturer dependant – the manufacturer goes out of > business, so would your solution. And that differs from a vendor-supplied, on-premise solution how? :^) Or are we now discussing "build versus buy"? > And the obvious advantages and disadvantages of an on premise solution I > would percept are: > Advantages: > 1) Onsite administration – You could do what you want, when you want, > how you want to the system as you please with no rules whatsoever apart from > system limitations I think you may have the wrong impression about SaaS, or at least in how it applies in the ITSM space. Not mentioning any names, but some are more flexible than others in this area. But again, it is all about the risk you are willing to accept that a patch from the vendor will break you, or that your upgrade to the next version will be long and painful. I am at a customer and the trip from 7.1 to 7.6 to 8.1 means leaving data behind twice. And that is low customization and on-premise. > 2) You can choose when to update if at all or stay on whatever version > works for you as long as you wish to. Lowers user training costs to a certain > extent. Again, without mentioning any names, some vendors provide you these options with SaaS. But just as BMC wants you to move along with the newer versions (and will cut off support, or ask for more support dollars if you don't) so do SaaS vendors. > 3) Your data is as secure as you want it to be. I would say that it is as secure as you can afford it to be. Most companies are simply not considered interesting enough targets to get owned by criminals and hackers. Also, many companies will not volunteer that they got hacked, especially if they lose no data that might be affected by the increasing number of privacy laws. > 4) Your solution life lasts beyond the manufacturers – if they go out > of business, you can continue to run their solutions for a while until you > have a better solution. Now I really think you are talking about build versus buy and not a vendor-supplied product. Granted, if you have downloaded the software and can re-install it you don't go down the day the company folds. But you are still racing against time. The only companies that I can think of that would continue to run with software where the vendor is gone are probably running their accounting software on an old Macintosh running Hypercard 20 years after the death of the product. (I only mention that because I met such a customer once ...) > Disadvantages: > 1) Onsite administration – You usually face higher maintenance and > running costs. > 2) You risk downtime during maintenance or upgrades or bug fixes even > with a good rollback strategy. These all seem to be the reverse of the above. > Any other advantages and disadvantages to the two strategies that I may have > not listed here? How about cost? I read an amazing statistic – I still do not know how to verify its validity – on the average cost per user per month by ITSM product and SaaS offerings (including BMC's) came out looking pretty good compared to on-premise. Remember, the cost of SaaS is pretty much an all-inclusive price (infrastructure, security, database, etc.) so most people don't account for those costs as they are already sunk in their organizations. How about allowing the business to focus on their core competency? SaaS usually also means PaaS and IaaS come along with it. As with data security, network reliability and systems management skill sets come at high costs. Part of the cloud concept is the move to data centers, where highly specialized skill sets can cover much larger computing operations. If your company makes machine parts you might not see providing world-class services in those areas as part of your core business. It might just make sense to let someone else do it better, cheaper. How about scaling computing resources to need and paying only for what you consume? Granted, in the ITSM space this is still a pipe dream, but true cloud computing (versus simple virtualization) promises that as demand goes down – say, during the evening, weekend, and holiday hours – the resources allocated contract. As the intent is to pay only for the computing resources that you use, this automatic contraction (and expansion, when demand is unusually high) can result in significant cost savings. But again, for our space that just isn't there yet. We are not really "in the cloud", but simply virtualized. Remedyforce comes the closest however, as it is a fully multi-tenant model, unlike all of the other offerings I have looked at. (I suspect we will start to see interesting things from that quarter in a year or so.) Just my $0.02. Regards, Dale Hurtt _______________________________________________________________________________ UNSUBSCRIBE or access ARSlist Archives at www.arslist.org "Where the Answers Are, and have been for 20 years"
