As an average consumer, I love the Cloud. The internet Cloud I mean. I have quickly become used to the idea that any photo I take on my iPhone, automatically shows up in Apple’s iCloud and on my iPad! I can put any document in my Dropbox account, and it is simply THERE, on all my devices including my PC.
I don’t have to worry about backups, or moving data around as I get new devices, and best of all, I can have all this for free! Well, at least that is how it feels. In reality, there are plenty of things to worry about, as explained here: http://www.techradar.com/us/news/internet/data-privacy-how-safe-is-your-data-in-the-cloud--1170332, but this is not what I really want to talk about today.
As a technology professional, I think I love the Cloud too, or perhaps I just love the idea of it. There are a ton of opinions and articles out on the web extolling its virtues. For example, who would not want 1.7 times greater ROI with the Cloud vs. an on-premise implementation? (http://cloudblog.salesforce.com/2012/09/cloud-and-roi.html). This all sounds wonderful, but I find it difficult to relate the marvelous promises to my everyday reality. It does not help that every vendor and pundit comes up with their own slightly different definitions of things, and I end up with more questions than answers. For example:
- I understand Cloud storage from my consumer experience, but what exactly are Cloud based software applications?
- What is the difference between a Cloud application and SaaS (“Software as a Service”)?
- How will I move data between a Cloud application and another application (either on-premise, or in the Cloud with another provider)?
It turns out that depending on what you are trying to do, the questions (and answers) will vary.
So here’s what I really want to talk about today: what does it really mean to an organization to deploy IBM Cognos TM1 on Cloud? This is a very real and specific situation, to which we ought to be able to give real and meaningful answers.
First some basics: TM1 on Cloud is a brand new offering from IBM, introduced in late 2013. The first Cloud release runs on TM1 10.2. You can read what IBM says about it here: http://www-01.ibm.com/common/ssi/cgi-bin/ssialias?infotype=AN&subtype=CA&htmlfid=897/ENUS213-358&appname=usn
Now, let’s do some analysis.
IBM equates TM1 on Cloud with SaaS (“Software as a Service”). A little bit of Google research will quickly tell you that not everyone would equate the terms “Cloud” and “SaaS”, although they are closely related and it is true that in both cases you are running your TM1 server on a computer in the cloud, and not in your own data center.
The practical benefit is that you don’t have to buy hardware, or request a Virtual Machine (“VM”) from your IT department, which saves you both money and time. You also don’t need to install the TM1 server software, which would also save you some money and time, especially if you were going to hire consultants to do it.
My favorite way to think about the difference between “Cloud” and “SaaS” is that the Cloud is the platform on which the software (as a service) runs. This article articulates it well: http://www.accountingweb.com/topic/technology/cloud-computing-versus-software-service
You can’t have “SaaS” without the Cloud, but you can certainly have the Cloud – which essentially is the whole internet – on its own. If you use Google Gmail, you are using SaaS. Salesforce.com – which helps you manage your customer relationships - is probably the most famous SaaS product used by businesses, and one of the largest, as you can see here: http://www.pwc.com/gx/en/technology/publications/global-software-100-leaders/saas-trends.jhtml.
So is TM1 on Cloud just like Google Gmail, or Salesforce.com? It actually has a lot in common with Salesforce.com, in that your license fee buys you access to a platform on which you can develop, usually with some outside help, a software application that addresses a business need. With Salesforce.com that business need relates to giving your sales organization tools for managing relationships and deals, while with TM1 it typically (but not always!) relates to financial reporting, planning, forecasting or consolidations.
Unlike Gmail, which you can use as soon as you sign up, both TM1 and Salesforce.com will not be useful until you have designed and implemented an application, a process that can take weeks or even months. The cost and effort of designing and implementing an application is the same, irrespective of whether you have TM1 on the Cloud or on a server in your own data center. Similarly the effort and cost to implement Salesforce.com is likely to be comparable to that of deploying an on-premise CRM solution from Oracle or Microsoft.
This is important to understand: just because something is offered on the Cloud, does not automatically mean that it does what you need out of the box. It would be nice if it did, but if you think about it, how could it? There is no way for Salesforce.com, or TM1, to know exactly how you want to run your business! In contrast, it is much easier for Google to figure out what you want from your email system, and to some extent they are even able to dictate what that is.
Now let’s talk about data flows. TM1, likes Salesforce.com, is at its core, a place where data is stored. Much of that data is typically extracted from other systems, for example a general ledger or ERP system. If you have TM1 on-premise, you have the option of loading data via an ODBC connection, or via text files. If you use an ODBC connection, TM1 can pull the data that it needs at any time from another system in your data center, and even on an automatic schedule. If you use text files, automation is also possible, but it is more complicated as there has to be an automated mechanism to generate the text files, and a system to coordinate the timing and order of events.
So how does this work with TM1 on Cloud? In the first (current) release of TM1 on Cloud, there is no possibility of setting up an ODBC connection, so the only option for loading bulk quantities of data is via text file. Furthermore the text files need to be transferred up to a specific computer in the Cloud via FTP before they will be visible to the Cloud TM1 server. IBM does plan to eventually offer options to make ODBC connections over the cloud, and the technology certainly exists, but for the foreseeable future one would need to design a business process around flat file extracts and FTP uploads. These things can be scripted, so automation is an option, but it will add cost to an implementation.
The next question you may ask is how does an end user interact with TM1 on Cloud? Well it turns out that TM1 on Cloud is a fully functional regular TM1 Server hosted in an IBM datacenter, together with another computer with the TM1 Perspectives client on it that you can log in to remotely over RDP and access to some disk space. This means that all the TM1 client tools that you may use with an on-premise TM1 server including TM1 Web, Cognos Insight and the new Excel CAFÉ client (watch a demo at http://youtu.be/4abO-5snYWI) are still available from each individual end user’s computer.
The primary development and administrative tool will continue to be TM1 Perspectives, which is accessed via RDP as explained above. The new Performance Modeler client tool is also available, but for now only Perspectives has the full range of capabilities needed to build a complex, highly tuned model.
So let’s summarize the fundamental realities of TM1 on Cloud:
- The biggest benefit is the startup cost and maintenance savings of not having to deal with hardware, installations and upgrades (more on upgrades to come!)
- One disadvantage is that you will not be able to suck data directly out of a source system (like a GL or ERP) using ODBC (but we are told this will be available one day)
- From a technology point of view, it is still the same powerful TM1
- There is no change to how implementations are done, or how much they cost
In this blog post, we have focused on the basics, but there is much more to discuss! Look out for upcoming blog posts that will expand on questions like data security, subtleties of licensing and why you might not want IBM to control the software upgrade schedule. We also plan to elaborate more on a bigger question: does TM1 on Cloud qualify as SaaS in the same way that Salesforce.com is considered SaaS?