New Report From Lustratus Research: A Competitive Review of SOA Appliances

We’ve just posted a new report up at www.lustratus.com that looks at a number of vendors offering devices that promise to lessen the burden of managing, accelerating and securing a service oriented architecture.

Authored by my colleague Steve Craggs, the report looks at “appliances” including IBM’s DataPower range, Layer 7’s SecureSpan XML Appliance and Intel’s SOA Expressway.  I say “appliances” in inverted commas because Intel’s product is wonderfully described as a software “appliance”. Surely the award for the most spin in a product category goes to Intel.

The document can be found here. Steve tells me that he’s planning to look at devices from the likes of Solace Systems and TIBCO in future reports.

Danny Goodall

Posted in All Blog Categories, Competition and Competitive Intelligence, research, SOA and Integration, Vendor-Related Posts and tagged , , , , , , , , , , , , .

6 Comments

  1. Hello There – It seems that this is a very provocative report, especially with respect to the statements made regarding the Intel product.

    First, off I have to say that I am from Intel, so as you must, please take my comment with a grain of salt.

    I hope, however, that the analyst does not confuse and equivocate a nuanced product available in multiple form factors for different usage models with “marketing spin.” The facts speak differently in this case.

    In fact, the Intel(R) SOA Expressway product (like some of its competitors) is available in three form factors (hardware, software and virtual image) – each of which can be properly called an appliance.

    “Appliance” here does not necessarily reflect a strict category of hardware only, but instead set of management and monitoring capabilities such as a real-time dashboard, self-healing capabilities, alarms, alerts, management clustering, and high availability with a familiar web-based interface and easy management.

    It is these capabilities that primarily characterize a software appliance. In this case we can think of a hardware appliance and then subtract out the physical security features. It is only natural that we can take this same form factor and package it for a virtual machine and we will arrive a similar form factor designed for a virtual private cloud. Incidentally, this is something especially difficult for a product only available as a pure hardware appliance.

    Finally, because the Intel product relies primarily on a software layer that performs machine language processing of XML, the addition of hardware adds only physical security prowess, and is not a necessary form factor for a high performance deployment. All in all, the product is truly available in all three form factors – no spin required. Perhaps some of these facts can “spin” the customer closer to the truth about this particular product.

    Blake Dournaee

    • Thanks for your comment Blake. This is a marketing-focused blog that looks at different vendors’ marketing strategies. So “spin” isn’t necessarily a bad word in this parish. In fact my “award” and “wonderfully described” comments in the original post were truly meant as a back-handed compliment. Either way my comments or views have no bearing on the report content. That is controlled by the report author Steve Craggs.

      It would be my guess that most people when confronted with the term appliance would expect to be talking about a device designed to perform a specialised task and many more may assume it involved specialised electronics. That software plays such a significant role in your overall offering, and the fact that you describe the product as a “software appliance”, is stretching the generally held definition of the appliance term in my opinion. Hence my comment about spin in a product category.

      But again please don’t think that this is a negative view on what you’re doing there. On the contrary I’m genuinely impressed with your marketing strategy for the product. Thanks once again for the clarifications and I’m happy to include them here. I’ll be sure to pass your comments on to the report author Steve Craggs.

      Danny Goodall

  2. Hi,

    I’d very much like to develop a point made in the report where it is asserted that software can never be as fast as hardware. On the face of it this is true but there’s the question of how you get the job to the specialist hardware that is going to do it’s work quickly.

    With crypto and XML operations there’s an overhead involved in moving the message to and from the acceleration hardware which means that quite often it’s just quicker to do the processing on the CPU where the message contents are already in the cache, ready to go. The end result is that for most real world messaging scenarios well written software can be quicker.

    This is particularly true when messaging concurrency is high. The fact that most hardware accelerators are single threaded does not work in their favour. There is much less of an impact on generalised CPUs which usually have eight hyper-threaded cores per server giving sixteen concurrent threads. For customers who do have use cases where crypto acceleration does work more quickly than leaving it to the processor, SOA Expressway does support crypto cards.

    Pete Logan, again from Intel.

    • Pete,

      Thanks for your comments, and those of your colleague. I think Danny has answered the ‘marketing spin’ question, so I will just deal with the performance one.

      You make some good points – of course any solution has to focus on end-to-end considerations, and certainly there is a cost with passing work off to another processor. However, the fact remains that I can always execute something faster in hardware than software – this is just a fact of physics. The question therefore becomes one of the balance between executing locally or remotely and using hardward and/or software. A lot of trade-offs relate to the proximity of the data, since obviously if a piece of processing requires frequent data references, then if that data is not local there is a high cost in constantly retrieving it.

      However in this case, the data issue is not a major factor. Take the example of XML processing – this is all about manipulating a data stream. As long as the maps to control that manipulation are local, then the execution time to carry out the XML transformations is far greater than the overhead of passing info to and from an appliance. So in this particular case given the criteria involved, I would contend that it would always be possible to drive higher performance using a hardware approach. Similarly with crypto – passing the data stream to be encrypted does not take much effort, because in the overall task of encrypting the stream the transfer cost is small compared to the effort required to perform the various encryption activities.

      Horses for courses as we say in the UK.

  3. Pingback: Truth Denied? | lustratus REPAMA

  4. I realize I’m responding to an inactive 2yro post, but perhaps some other wayward information googler will stumble across this as I have and benefit from the discussion. Firstly, I’m an agnostic in the discussion of SAO specific appliances, software or hardware varieties. That said, it appears there’s been a healthy dose of kool-aide being passed around with regard to DataPower and asic acceleration.

    True asics are faster than software. True also is the fact certain aspects of XML parsing are faster on DP than on the Intel ESG. However, as soon as you start doing anything more complex than scheme validation such as XML signing and encryption the message traffic is punted to the CPU and is software processed. Once that happens it becomes a crushing defeat to the whole asic conversation. Becasue of this design constraint IBM positions themselves inside the infrastructure closer to mainframes and apps servers to perform protocol and message mitigation tasks rather than on the edge in a DMZ.

    So the marketing “spin” award I think belongs to IBM.

    BTW – Just try to get the actual hardware stats (CPU, Mem, data bus, etc.) out of IBM for their product. It’s been my experience they won’t release it. Also, the DP product is run by the websphere BU, not the hardware devision.

Leave a Reply

Your email address will not be published. Required fields are marked *