Network Sizing: Assumptions or Well Informed Decisions?

In an earlier post, I mentioned I would do some more work on the Exchange Client Bandwidth Calculator from Microsoft.

I will try and explain why I strongly disagree with these kinds of tools.

Each Domino server logs the activity performed by end users. For each user session, an entry is written in the log that records start and stop time (and duration) of the session along with the network bytes sent and received by the client from the server. There's more data logged, but this is not relevant for this article.

Because DNA collects all this data over a 7 day period, we can aggregate this session data over all activities from all end users for each DNA customer:

• kbps_rcvd = 8 x sum (bytes_received) / 1024
• kbps_sent = 8 x sum (bytes_sent) / 1024

Doing so for the most recent 100 DNA engagements, shows the average score for each customer in the graph below:

Both axes are scaled logarithmically for visualization purposes, and the overall DNA average is shown by the cross-hair. Each icon in the graph represents a customer organization, the larger the icon, the more users the organization has (seat counts range from 10 to 93,000).

Two things we can observe from these customer scores:

1. Average Client Network Bandwidth Consumption ranges from 1 to 49 kbps per user across these 100 recent DNA Customers;
2. Overall DNA Average equals 10.03 (received by users) and 2.40 kbps (sent by users);

Let's take a look at what Microsoft is doing in their calculation workbook: 

1. In the workbook (worksheet named 'Data Tables') they show peak traffic levels in kbps for each type of client
2. Microsoft assumes a normal distribution when it comes to User Demand on a typical working day

Using the default profiles in the workbook, this would result in per user averages of 6.36 kbps received for Very Heavy Users and 1.42 kbps received for Light Users. Applying this to our Real World Customers, clearly shows how this would cause significant under-capacity in 70 of our 100 recent customers:

What's causing the huge differences?

I think Microsoft makes a big mistake assuming that all users work along a normal distribution, as shown in their graph (worksheet 'Tools'):

The reality however is completely different, in several ways:

A.

Analyzing the distinct number of hours that users are active for my 100 recent customers, shows that 20% of all users are active less than two hours per day (blue line). 40% of all users are active no more than 5 hours per day.

So this means that a normal distribution -as Microsoft is assuming in their workbook- is not realistic. In fact,  the analysis of the workload for one of my customers illustrates why Microsoft is wrong:

Notice how most users are online between 7 and 17, but the network load is very much concentrated around the morning hours. This is when the remote workers (that are typically online 1-3 hours per day) consume all their new mail, especially on Monday morning.

B.

Having experience with large customers, Microsoft should realize that users can be located in different time zones. So unless you place a data center in each time zone, and have all users from that time zone connect only to that data center in their on time zone, you will see demand patterns that do not match a normal distribution. Instead you will see multiple distributions come together in one data center. 

Finally, let me show you the real world: kbps received and sent for 928,420 Lotus Notes Users  (I love Tableau Software...):

Dear Microsoft, please beware that the real distribution for network bandwidth consumption of end users lays as much as 8 decades (10^8) apart. Each customer is a unique. Even within each customer you should analyze the Real End User Demand in each office location, before calculating network requirements.

Apple TV 3, a first impression

Looking for a way to easily stream my Africa pics to my TV, I stumbled upon the Apple TV and decided to buy it for EUR 109. Package comes in a box and the form factor of the gadget is small & sexy, black matches my TV furniture ;-)

Installation equally simple, had it up & running in 5 minutes. Beware to have both an HDMI and optical audio cable available.

Currently, I have about 12,500 pics to view. In several folders. Obviously my Africa 2011 folder is the largest, with some 2,500 pics shot from a Canon 550D at 18M. 

I took me a while to find out that you don't push pics to the Apple TV, as you would expect with Airplay. Instead, you make the Apple TV pull the pics from your iTines library. Had I known sooner, I wouldn't have bought the iPhoto software at EUR 12,00 from the Apple Store :(

Here's the nice discoveries I made:

1. you make iTunes share your photo library

2. you tell the Apple TV to look at Computers and simply select the iTunes library you just shared;

3. you configure the Slideshow in the Apple TV and simply watch your pictures

4. be patient, coz the systems may need some time to sync. (this seemed to take up to a day in my case)

What's even nicer, I didn't configure music in my slide show. I simply use Spotify off my iPhone to push music into my Apple TV. Discovered that pausing the slide show with the Apple TV Remote also pauses Spotify on my iPhone.

Of course the Apple TV has more features, but I'm sure others have investigated those in far more detail than I do. If you have any questions about photo-streaming with Apple TV, feel free to reach out though.

My conclusion: I like it, and am still happy I abandoned Microsoft earlier this year.

Exchange Client Network Bandwidth Calculator

Last night one of my friends in IBM asked my opinion about the Exchange Client Network Bandwidth Calculator (it's here). I gave him my response, after I had downloaded and played with it. Here it is.

Since 10 years, my work and passion is all about analyzing network traffic for large Lotus Domino customers, I have detailed network and usage statistics for more than a million Notes users in my DNA data warehouse.  And I maintain benchmarks to compare customers against. So I think I can judge.

From a first look and play with this 2012 calculator spreadsheet, it looks pretty complete in the sense that Microsoft have incorporated a lot of variables to represent end user behavior. Now let's see what happens if we run a couple of scenario's:

Scenario A: 10,000 users with a Medium Profile

1. Leave all parameters on their initial default values in the worksheet named 'Input';
2. In the worksheet named 'Client Mix', specify all users to have a Medium Profile at each site and in such a manner that every version and type of Outlook is assigned 10,000 users;
3. Add a column to the right where you divide the resulting bandwidth by the number of users, thus calculating the kilobits per user: (=1024*Table47[@[Network Bandwidth )

Click to enlarge

Recommended network bandwidth is appr. 4 kilobits per second, per user.


Scenario B: 10,000 users with a Very Heavy Profile

1. Again leave the parameters to their initial values;
2. Change the Site User Profile to 'Very Heavy';

Click to enlarge

Recommended network bandwidth ranges from 6.5 to 11.75 kilobits per second, per user.

My observation:

Let's apply the Microsoft Recommendations to the network bandwidth requirements I reported to an organization with 9,500 users and 101 office locations world wide.

Click to enlarge

The blue bars are network demand levels observed over a 7 day period, as a consequence of measuring the real end user demand in those locations. The two red lines come from Microsoft's estimates for Medium and Very Heavy user profiles, that I applied to the user count in each of the office locations.

Of the 101 office locations, Microsoft's estimations would result in a severe shortage of network capacity for 6 locations, even when applying Microsoft's most heavy profile estimates.

Would we follow Microsoft's Calculator for Medium Profiles, more than 50% of all office locations would end up with severe shortage in their network capacity. For the remaining locations, the customer would end up with significant over capacity (56% to be exact, averaged across all sites).

Why is the Microsoft Calculator giving wrong estimates for almost every site?
Do the default user profiles not represent the behavior of users at my customer properly? Do the users in my customer send and receive more messages, and are these messages perhaps smaller or larger than what Microsoft put in their profiles? I don't think that matters. Putting more load per user profile would lead to even more over capacity at half of the sites, while still leading to shortage in others.

The real problem with this calculator is putting people into profiles. You cannot predict end user behavior, you can only observe what they are doing today. It's ok to classify your observations into profiles and categories to e.g. present pretty pictures. But it is very dangerous to use those classifications and generalize entire user populations for sizing purposes.

My Conclusion:

Predicting network requirements will almost always result in either over-capacity (waste of money) or under-capacity (end users complaining about poor performance).

With DNA, we measure the Real End User Demand over a relevant 7 day period. It is far safer to expect that end users will show the same behavior tomorrow, as we observe today, than to apply an assumption that 4 or so profiles fit all your end users.

My Recommendation:

If your sole objective is to migrate away from Domino, regardless the consequences and as quickly as possible, go ahead and use Microsoft's Network Bandwidth Calculator. You'll love it.

If you prefer to make well-informed, fact based decisions with regards to your platform strategy, based on the Real End User Demand, ask IBM to perform a detailed study on the end user demand in your organization.

Side Note:
I am not an expert in Microsoft Exchange or Outlook. Although theoretically it could be possible that network consumption of outlook users is significantly lower than a typical Notes user, I doubt that this is the case.

Cooking diner, the old fashioned 1912 style

Today I am cooking a very traditional Dutch dish called Boerenkool (Cow Cabbage or Tree Cabbage?) with Rookworst (smoked sausage), using the cookbook my grandmother Gerda Smit-Goedhart had in use when she was attending the Amsterdamsche Huishoudschool in Amsterdam around 1912. 100 years ago.

BERICHT BIJ DEN VIJFDEN DRUK
Enkele recepten werden ingelascht, op nieuwe vindingen werd de aandacht gevestigd, het gebruik van boter en eieren ingekrompen of door andere dingen vervangen, waardoor ik hoop, dat het boek aan bruikbaarheid gewonnen heeft.
C.J. WANNEE, voorheen leerares in Koken en Voedingsleer aan Die Inrichting.

The fact she describes how the use of eggs and butter was reduced to a minimum in this book-print, makes me wonder to what extend there was a shortage of these ingredients back in 1912. Was there a birth-flu and mad-cow disease happening one hundred years ago?

Reading the introduction, my eyes were caught by the following pages: How to set the table and how to serve dishes, the old-fashioned proper way.

Let me translate a couple of bullets from these pages:

1st
Extreme sanity should be in place, both on material that we will be using, personally with regards to the ones who are going to serve the dishes, as well as on the surrounding (table) where the meal is being consumed.

3rd
Rest and silence while serving, so that it does not only liberates the host from all her worries, but also positively surprises guests. One should pay particular attention to placing the glasses and plates in a silent manner, walk in a manner one cannot hear and not distract the attention of the host and hostess.

This book is about old-school rules for old fashioned dinner pleasure. A hard thing to find nowadays...

Anyway, here's the Dutch instruction for the Boerenkool dish (literally translated as Farmers Cabbage with Patatoes and Smoked Sausage):

How come this took 30-35 minutes to cook in water? Nowadays, 15-20 minutes are sufficient. Could this be the influence of today's crop manipulation?

Notes Workspace on BlackBerry

Go check this out: Teamstudio showed me a product today that brings your Notes apps to mobiles (Black Berry, iPhone, iPad, Android).

Here's a link: http://unplugged.teamstudio.com/