- In extensive Fiori testing, the first thing we observed is how slow the interface is.
- This caused us to perform a speed test which we published here.
The Fiori Cloud is a strange introduction by SAP. You will learn about the Fiori Cloud and how accurate the claims for the Fiori Cloud are.
What is the Fiori Cloud
The Fiori Cloud is one of those strange artifacts that SAP brought out a while ago. The Fiori Cloud is a bit confusing because Fiori is just a UI, so you can’t have just a UI that is in the cloud. It has to be connected to an application layer and a database.
However, what the Fiori Cloud really is, is an online demonstration of Fiori with S/4HANA. Upon investigating this, we found something peculiar, which is related to the Fiori’s Cloud’s speed, which is the topic of this article.
Poking Around The Fiori Cloud
The Fiori Cloud is easy to access.
Once you get into it, it brings up the well-recognized Fiori tiles or squares.
The Fiori “tiles” is sort of the opposite of the SAPGUI, which is driven by transactions, or by navigating a very large tree structure.
With Fiori, the squares are selected to get into each transaction or screen. This demos nicely, but there are questions related to how well this design scales.
But Fiori has a nice search feature. This takes you directly to the item or the right square.
Once you select the item you want, often from a number of options that all meet the search criteria, you can be taken into the item or square. It has a very nice feel. But it is unclear to us if is an efficient method. It greatly depends upon the search function working, which we are about to dive into.
The item you highlight points to the right square which you can then select.
Fiori’s Hit and Miss Search
The search sometimes works great and is quite fast when it does work. But the search does not always work.
But once you select the item, this is a common response.
There is a square called Working Capital Analysis. But where is it in this search? It should have come up on the right as an option once any of the keywords were typed in. It was there one time we logged in, we know because we wrote down the time it took to open, but it disappeared the next time we logged in. That is a first.
This repeatedly occurred when we tested different searchers. Some words worked, but others didn’t. But while the search worked intermittently, Working Capital Analysis was the only square to just disappear from the UI. We checked by scrolling rather than using the search.
How does that happen?
The Best UI in Enterprise Software?
SAP has been carrying on about Fiori as the future. Hasso Plattner called it the best UI in enterprise software. But then why isn’t something basic like this fixed?
If there are many squared (not the 20 or so for the demo) how is the user supposed to find the square? Scroll the entire list of thousands of squares? That is not a feasible option.
SAP proposes that Fiori will eventually make the SAPGUI obsolete. That is not going to happen with the search still not working; squares decide to disappear, combined with such a small amount of coverage over SAP. We covered that second topic in the article The Strange Changes with the Count of Fiori Apps.
Why are we the only ones to publish on this topic? The Fiori Cloud is available for anyone to go and check and test. But as we have pointed out in previous articles, all the money in consulting and IT media is in agreeing with whatever SAP says. Fact checking is simply not a focus. Even if Deloitte or CIO looked into this, they would never publish on their findings.
The entirety of the information apparatus that covers SAP is there to promote SAP, not to fact check SAP or to tell their clients and readers the real story on SAP.
It is exceptionally difficult to get any SAP consultant who will tell companies the truth about SAP. Most SAP consultants value their relationship with SAP and with other SAP consultants more than they do their relationship with their clients. Lying is rampant in SAP consulting. The objective is to make SAP look as good as possible, the truth is considered only within the context of a massaged narrative.
Speed Tests for Fiori
After we got through the search problem, we were struck by how often we kept seeing the wait page for Fiori that looks like this.
We found this latency issue at several different locations, and therefore different Internet speeds.
We did not notice any other latency issues using any other website that we accessed at these same locations. We checked the speed at one location with Speed Test, and here are the results.
So this was not a perfect Internet connection, but it was better than average, scoring four out of a possible five stars.
When we found an even faster connection, one with five stars, we found that the Adjust Stock square/transaction took 4.49 seconds to open, 2.23 seconds longer than when tested at the slower Internet location (with four stars rather than five).
This slowness of Fiori is not a function of the Internet connection, its a function of the Fiori server, database, etc..)
The following is how long it took to simply get into the transaction screen by selecting the initial screen.
|Fiori Transaction||Load Speed in Seconds (Test 1)||Load Speed in Seconds (Test 2)|
|Track Sales Orders||2.793||4.25|
|Global Cash Position||3.85||4.15|
|Working Capital Analysis||12.61||N/A (Square Disappeared so We Could not Retest)|
*All timings were taken using an Android stopwatch app.
What About The Effect of HANA?
The presentation of HANA has been that it would enormously speed both analytics and transaction process. Hasso Plattner has stated that HANA will deliver zero latency to all applications. If we take Hasso Plattner at his word, this means that the Fiori Cloud squares/transactions should have been limited only in the Internet connection latency, as the web, database and application server should have performed an instantaneous return. The total number of seconds should have been 0.00, exclusive of the Internet time.
We tested the fastest web page we know of which is Google at .486 seconds. But Google only returns text (we tested it searching for a word). Still, this would seem to be the rough latency of the Internet itself, .486, or roughly 1/2 of a second. So while Google is very close to zero latency, SAP is far off the reservation.
*At 1/2 a second, as one has to hit the return button and move one’s finger to hit the timer, verify the data populating the web page, and then re-hit the timer a very accurate measurement is not possible.
However the Fiori Cloud undoubtedly has HANA, yet the application transactions take an average of 4.33 seconds to load.
Fiori Versus Our Website?
As a means of comparison, we checked the download time of one of our own web pages at Brightwork Research & Analysis. We end up with a time of 3.09. However, our pages have images on them, which means the page is larger than the Fiori pages that are being rendered. The speed will depend upon how many images the web page has (primarily). We do have pages that will render more slowly than 4.33 seconds (the Fiori average), but this is a function of the having quite a few images. Furthermore, we have far more text as well as formatting in a single page than in any Fiori screen that we tested.
SAP Fiori + HANA Losing to Open Source Products?
However, why are our larger web pages loading faster than the smaller Fiori pages that are only rendering numbers and text? Are we using some super fast backend? Hardly. We like our web host, but it is no top end setup. If we wanted to invest more money per month, we could get the speed faster, at a quite small cost. We could for instance move towards a dedicated server at our current host. That would increase the hardware available to fulfill requests.
What about the database? Is an advanced top-end database that the secret to our performance? Nope. Our web host uses MySQL. MySQL is owned by Oracle, but it is an open source database. MySQL is free. Does MySQL have a column data store and “in memory architecture” as does HANA? Nope. In fact, HANA does not compete with MySQL.
SAP has stated that HANA is faster than any of the top end databases offered by Oracle, IBM or Microsoft. But they are certainly not referring to open source database projects. Open source databases like MySQL, MariaDB, and PostgreSQL are not even part of the HANA conversation.
How about the application server? Must we be using a space-age application server right?
Our host uses Apache. Once again, Apache is an open source project and is free. Fiori uses the SAP Fiori Front-end Server. It is based on a NetWeaver Applications Server ABAP.
We did not start out trying to illustrate that the Fiori Cloud is slow, or that the Fiori transaction search only works inconsistently. We discovered these while just taking the Fiori Cloud for a demo. We have spent a lot of time analyzing HANA and Fiori which we have covered in articles like What is the Actual Performance of HANA? and What is in the Fiori Box. Even in a basic analysis, like this one, we find that SAP’s claims regarding Fiori and HANA do not check out. If SAP’s “in memory architecture” is so great, why are Fiori and HANA outperformed by our combination of WordPress, Apache, and MySQL — all of which are open source and free products?
The issue that we see is that no one is fact checking SAP and publishing the results. Therefore SAP marketing is sitting there proposing a virtually unlimited number claims which go unchallenged. If SAP’s claims were true, then it would be annihilating an open sources configuration that is common for the vast majority of websites. But it doesn’t, it fact, it loses to it.
Furthermore, the Fiori Cloud is supposed to be a showcase to demonstrate how superior Fiori is, and with SAP’s virtually unlimited resources it should be configured for speed. Oh, and the search box should work, and it should work 100% of the time and without disappearing transactions.
Financial Bias Disclosure
This article and no other article on the Brightwork website is paid for by a software vendor, including Oracle and SAP. Brightwork does offer competitive intelligence work to vendors as part of its business, but no published research or articles are written with any financial consideration. As part of Brightwork’s commitment to publishing independent, unbiased research, the company’s business model is driven by consulting services; no paid media placements are accepted.