No more hiding in the server room
Resisting SaaS For IT Management? Bad Career Move, Dinosaur
The question is very simple: are you going to be dragged by the business people in your organization to introduce SaaS solutions into your IT management environment, or are you going to lead your organization to it?
The Extinction Has Begun
In the past 10 years, SaaS has become a mainstream delivery model for many business applications such as Customer Relationship Management (Salesforce.com, Oracle On-Demand), Email/Calendar/Collaboration (Google Apps for your domain), and Operating Systems as a service (Microsoft Azure). SaaS solutions have also been introduced in the IT Management domain, with thousands of IT professionals leveraging SaaS to be able to spend less time on the ‘plumbing’ of IT (e.g the infrastructure layer) and spend more time on delivering a service to the different business units.
* Online backup (Mozy, IronMountain,IBM)
* Desktop Management (Kaseya)
* Email Archiving/security (Google/Postini,Dell/MessageOne)
More examples can be found on Jeff Kaplan’s SaaS Showplace.
There are many benefits of SaaS over on-premise solutions, including but not limited to:
Lower Cost – Most SaaS solutions are offered at a lower subscription cost (Monthly/Quarterly/Yearly) with no up-front fees, compared to on-premise solutions. Much of the savings are realized because of the SaaS vendor’s lower cost of sales and a lower cost of deployment. The subscription model means that funds are coming out of the operating expense budget and not from the capital expense one.
Simpler Solution – The less software and hardware components installed on-premise, the simpler & faster the solution deployment becomes. The time to value is much faster compared to on-premise solutions – usually measured in minutes.
Lower Risk – No shelfware! In a SaaS model the customer can unsubscribe or cancel the subscription, and with no databases and application servers to install, this means lower overall risk to IT and the business.
Faster Innovation – User community innovation is what drives most SaaS companies’ roadmaps. Unlike On-premise solutions,where new functionality is generally introduced twice a year, a SaaS Company, using an Agile development methodology can introduce new functionality every 6-8 weeks. Moreover, due to the complexity associated with upgrading on-premise solutions, most customers go through the upgrade only during scheduled maintenance windows. This means at any given time, the customer is anywhere from 9-12 months behind the latest and greatest functionality available.
SaaS Data Management applications simplify the traditional on-premise deployment process and eliminate the need for additional:
* Compute resources
* DB resources
* Storage Resources
* Agents
* Software changes / updates
* Backing-up / replicating the IT management App infrastructure itself…
So either SaaS your IT management or get off the pot. And here’s a list of recommended reading:
* The Big switch
* Who moved my cheese
* What killed the dinosaur
The Extinction Has Begun
In the past 10 years, SaaS has become a mainstream delivery model for many business applications such as Customer Relationship Management (Salesforce.com, Oracle On-Demand), Email/Calendar/Collaboration (Google Apps for your domain), and Operating Systems as a service (Microsoft Azure). SaaS solutions have also been introduced in the IT Management domain, with thousands of IT professionals leveraging SaaS to be able to spend less time on the ‘plumbing’ of IT (e.g the infrastructure layer) and spend more time on delivering a service to the different business units.
* Online backup (Mozy, IronMountain,IBM)
* Desktop Management (Kaseya)
* Email Archiving/security (Google/Postini,Dell/MessageOne)
More examples can be found on Jeff Kaplan’s SaaS Showplace.
There are many benefits of SaaS over on-premise solutions, including but not limited to:
Lower Cost – Most SaaS solutions are offered at a lower subscription cost (Monthly/Quarterly/Yearly) with no up-front fees, compared to on-premise solutions. Much of the savings are realized because of the SaaS vendor’s lower cost of sales and a lower cost of deployment. The subscription model means that funds are coming out of the operating expense budget and not from the capital expense one.
Simpler Solution – The less software and hardware components installed on-premise, the simpler & faster the solution deployment becomes. The time to value is much faster compared to on-premise solutions – usually measured in minutes.
Lower Risk – No shelfware! In a SaaS model the customer can unsubscribe or cancel the subscription, and with no databases and application servers to install, this means lower overall risk to IT and the business.
Faster Innovation – User community innovation is what drives most SaaS companies’ roadmaps. Unlike On-premise solutions,where new functionality is generally introduced twice a year, a SaaS Company, using an Agile development methodology can introduce new functionality every 6-8 weeks. Moreover, due to the complexity associated with upgrading on-premise solutions, most customers go through the upgrade only during scheduled maintenance windows. This means at any given time, the customer is anywhere from 9-12 months behind the latest and greatest functionality available.
SaaS Data Management applications simplify the traditional on-premise deployment process and eliminate the need for additional:
* Compute resources
* DB resources
* Storage Resources
* Agents
* Software changes / updates
* Backing-up / replicating the IT management App infrastructure itself…
So either SaaS your IT management or get off the pot. And here’s a list of recommended reading:
* The Big switch
* Who moved my cheese
* What killed the dinosaur
Views: 30
Replies to This Discussion
-
Permalink Reply by Wesley "Nonapeptide" on
-
I'm slightly skeptical of SaaS , mostly for flexibility issues but also some privacy issues. Specifically, I'd love it if every email system I dealt with was offsited (new verb!), but it can't happen sometimes because of the inherent inflexibility of many offsite mail solutions as well as a perceived lack of confidentiality. Then there's legal compliance issues that SaaS simply can't address until the laws change. All in all, I don't think SaaS is going to thin out this admin's datancenters very much in the next 5 or 10 years in spite of my desire to rid myself of in-house Exchange. *sigh*
-
Permalink Reply by Tsahy Shapsa on -
Wesley, you bring up a few good points:
- Flexibility - I've used several CRMs in my career. Maybe Siebel is very flexible, but my it took my company over a year to implement it. And because every change was painful, guess what - we used maybe %20 of the functionality. We're acheiving much more with our salesforce.com implementation today. And so are thousands of other companies. In many cloud services you
Privacy - There's no law that states one can't use Cloud / SaaS for privacy issues. Sensitive data needs to be 'well protected' (read: encrypted, user access controls in place etc). Thereare financial services companies consuming SaaS services, Healthcare ones are consuing online EMR services, Blackberry hosts/caches everybody's email on their servers and even the US government launched their own cloud service - https://apps.gov
This Admin's data center is already changing:Mission critical Applications such as CRM, Payroll, HR etc have alreadyfollowed Elvis and 'left the building'. The thing is that this evolution, in many places, isn't driven by IT, rather it's driven by end-users / business units. So what's an IT pro to do? one option is to resist the change. Another option is to embrace it. Yet another option is to be the change agent. Be the one in your company that introduces AWS and enables a new business initiative to come to fruition. Be the one in your company that's able to look at different cloud architectures and recommend the best fit for a specific project in your company. Be the one that tells your CIO / CFO how you can achieve financial efficiency using an infrastructure utility service where it makes sense, taking the burden off of your capital expenses and using operational ones instead (CFOs love this). Just as you do when it comes to electricity, just as you do when you trust your money to be hosted and managed by the banking system..
As of your exchange pains - my company uses Google Apps (along with 1 Million other businesses worldwide) and we love it.
-
-
Great points Tsahy! I'll continue to ruminate on them. The compliance issues are concerning that you have to rely on the vendor and hope that when they say they're compliant, they really are. And you have to negotiate compliance contracts in the first place. At least if I manage the infrastructure, I can peer into every operation of the compliance chain and verify that they are truly in compliance. Here's an article that outlines some of the issues regarding SaaS and compliance: http://searchcompliance.techtarget.com/tip/0,289483,sid195_gci13407...
It's not that it can't be done, just that it's not the silver bullet that some might have hoped for and has its own set of complexities.
One of the organizations that I worked for seriously looked into Google Apps for email, however there is currently one gaping hole in Google Apps and Outlook integration: No support for Outlook Tasks. My organization relied heavily on Outlook Tasks. As a result, we had to look at the more expensive option which was hosted Exchange where we basically had remote desktop access to an Exchange server that replicated with our AD. We managed the Exchange server ourself. Is it really SaaS if you manage the server? I'm not so sure. Maybe it's pseudo-Saas. It's still better than onsite Exchange, but there's still a significant amount of hands-on. At least we don't have to worry about the battery units, physical infrastructure, and the hosting environment took care of the failover and backups -- which is why I'm torn between calling it true SaaS and a SaaSy-colocation arrangement.
In the end, I see SaaS as mostly easing infrastructure requirements from out of my datacenter. Namely and in order of my perceived level of importance: less backup/DR headaches, less redundancy headaches (I don't have to implement some juggernaut of a failover cluster with offsite replication, etc.), less power worries, less space, less cooling.
Again, great points! Thanks for the thought provoking article!
-
Permalink Reply by Tsahy Shapsa on
-
Wesley, I'm not trying to argue that SaaS or any cloud technology for that matter is a 'silver bullet'. What I am arguing is that SaaS should be part of the portfolio of technology-based services IT provides back to the business - with that, all the 'good old' IT practices should remain in place. Reading your last sentence - I couldn't agree more!
-
-
Exactly. I think we're on the same page. I didn't think you were implying that SaaS was a "silver bullet", I was merely commenting that some people out there (usually sales agents for SaaS) companies seem to think that SaaS can do everything including walk my dog which isn't quite the case... yet. As it is I'm happy to report that some of my users are on basecamp and some use an event coordination SaaS tool. It's great for me. When they have a problem I just say "Hey, you bought it. You can call their tech support." =)
-
Permalink Reply by Robert Chipperfield on -
An interesting read - especially the discussion in the comments.
There are some SaaS type bits that I use and love - I suppose if you stretch the definition a little, how many companies use MSN Messenger for internal communication rather than deploying OCS? We certainly did a while back, unofficially at least. We use NTRsupport for remote customer support, and that's saved us huge amounts of effort and infrastructure - for the most part, it Just Works, and it's great.
But there is the risk of "silver bullet-itus". Particularly with the capex / opex debate: if something costs $1000 up front to buy and run in house, plus maybe $100 / year as a suitable share of a server's power, cooling and replacement costs (I'm assuming you're hosting multiple apps on a server here!), but you could get a SaaS equivalent for $50 / month, then after one year - great! You're $500 up! But in three years, you've spent $1800 on SaaS rather than $1300, and it only gets worse with time.
An interesting analogy I remember is the phase several large corporations (including the BBC) went through selling their building infrastructure off, and renting it back. Short term, it provided a huge cash injection, but long term, it's obviously going to be more expensive (otherwise, how can the company now owning it make a profit?). Similar with outsourcing call centres for tech support - how many companies are now bringing these back in-house?
As I started, I'm not denouncing all SaaS as a false economy useful to only those with visions as long as their next term of office; far from it. Clearly in some applications economies of scale mean it really can work out better to centralise on a huge scale.
Maybe it's more a plea to look beyond the next year's budget and take the long-term view?
(A brief post-amble: at Red Gate, we also needed to replace our CRM system recently. Of course we investigated the usual options, SaaS and otherwise. Our final answer? We wrote our own. It's awesome - for us. Probably useless for anyone else, and certainly more expensive than any of the other options in years 1-3. But long-term, it fits our needs perfectly, and I think that'll pay for itself many times over.)
-
Permalink Reply by Tsahy Shapsa on
-
Robert,
There are a few more variables that should go into the TCO comparison of yours, they basically have to do with the 'lost opportunity' of all the reources invested in implementing an in-house solution:
1. Alternative cost of money - that money NOT invested upfront, can be put to work on other projects and generate cash-flow
2. Additional HW & SW required to to support an in-house solution
3. Cost of IT time/resources required to implement & support an in-house solution. IT time could be spent on 'higher level' items, closer to the business, discussing architecture, requirements and providing a service, without worrying about the plumbing..
-
Permalink Reply by Robert Chipperfield on
-
Absolutely - my calculation was obviously a vast over-simplification.
But someone has to worry about the plumbing, and if it's not your company, then whoever it is will be charging a mark-up on whatever it costs them, so in order for it to work out cheaper in the long run, they need to be able to provide the service significantly more cheaply than you can.
Which is probably either through a large economy of scale (where it truly is win-win), reduced quality of service (common in off-shore call centres) or an unsustainable business model on their part (how will Skype make a profit in the end?). It's worth ensuring it's the first reason and not the second two.
-
-
My post below was being worked on as you posted this. Interesting point about Skype. They probably make money from shady syndicates pumping money into the business to keep the encryption strong and not put in any backdoors. Of course, having said that, I'll probably be found floating face down in a lake somewhere... =)
-
-
I'm a big proponent of "Time is Money", or rather "Time is more valuable than money". As Tsahy said, if a $500 a year SaaS solution is more expensive in dollar amounts after the second year, that's not necessarily a loss. Think of the time that I didn't have to restore that server after the server room had some "issues" with it's CRAC. Think of the time I didn't have to spend on migrating one more server when we moved datacenters. Think of the time not spent trying to expand my backup window. Think of... well, you get the idea.
EDIT: I forgot the biggest "Think about" plus for me. Think about the time not spent trying to scale the hardware that the software runs on up and down based on user demand. Of course, that's taken care of nicely if you use a virtualization scheme in-house, but still.
-
Permalink Reply by Robert Chipperfield on
-
Scaling is certainly a big benefit of SaaS (well, really of any centralised system). As is load balancing over time - chances are when you're using it, someone else isn't, so you get a much more steady load on the system as a whole.
On the other hand, though: think about when the service does go down, and all you can say is "it's a problem with the supplier", or worse, when your Internet connection goes down, and suddenly no-one can e-mail the person next to them. Extreme example of course, but worth considering for business critical systems. (I wonder how many places spend vast amounts on the aforementioned behemoths of failover clusters, but only have a single ISP providing the pipes into the building?)
I guess we're back to the silver bullet, or lack thereof: SaaS can be great in some places, but isn't the necessarily the solution to every problem.
EDIT: An aside on single points of failure: this made for some interesting reading, from the bottom: http://aaisp.blogspot.com/search?q=power
-
-
Indeed, as more things are offsited (Yes! Used it twice in one thread!) the internet connection becomes more and more vital. But I wonder if it was ever not vital? Being unable to email the person next to you seems trivial if you can drop the message in a shared folder. Being unable to send a quote to a potential client seems to be a bigger issue, one that would exist if that particular service was on or off site. Redundant local ISPs is less of a luxury these days, but I suspect it was never a luxury but was only perceived to be one.
You mention a very uncomfortable conversation that can come up with SaaS: "Why is it down?!", "Uhhh... it's out of our hands. They say they're working on it." Such a feeling of helplessness and vulnerability. However, I'd think that if you went with a trusted SaaS providor they'd have more of the resources necessary to fix a major outage. More dedicated people, more highly trained people, more money, etc. Yes, that is a huge assumption on my part. One that would have to be empirically proven to be trusted. Maybe that's just my coping strategy kicking in. =)
I think availability is low on the list of risks (assuming your SaaS is provided by someone other than "Bob's Bait and Tackle Emporium, with new Datacenter Offerings!" I'm still stuck on the inherent flexibility of SaaS. They'd better provide good APIs and allow the developers to create some customized forms and maybe even data types.
EDIT: I hope you're happy. That blog you posted is going to keep me up at night. =)
© 2012 Created by Elizabeth Ayer and Michael Francis.
Powered by
