Data operation
Unlike transactions with attributes of the data to determine the size of 8 MB. (The
size of a block of data, some NFS. More recently announced version of NFS
protocol + allows blocks of data up to 4 GB. But this has not changed the nature
of the data). Furthermore, while for each file, there is only one set of
attributes, the number of blocks up to 8 MB in a file can be large (potentially
reach several million). For most types NFS-serverov blocks of data are not
usually are cached, and thus servicing of requests associated with a significant
consumption of resources. In particular, to handle data requires much greater
bandwidth network : each of the data transmission includes six large packages
for Ethernet (two for FDDI). The possibility of overloading the network is a far
more important factor in the consideration of the data. Somewhat surprisingly,
but the majority of existing systems dominated transactions with attributes, and
not just in the data. If the client wants to use the system NFS file stored on a
remote file server, it gives the sequence of search (lookup) to determine the
location of the remote file directory hierarchy, followed by the operation
getattr masks for human access and other attributes of a file; Finally, the read
extracts the first 8 MB of data. For a typical file, which is located at a depth
of four or five levels of subdirectories remote hierarchy, simply opening the
file requires five or six of NFS. As most files rather short (average for the
majority of less than 16 MB) to read the entire file requires fewer operations
than for its search and opening. Recent studies of Sun found that from the
operating system, BSD 4.1 average file size has increased from about 1 MB to a
little more than 8 MB. To determine the correct server configuration NFS
primarily be attributed to one of two classes, in accordance with the dominant
workload for alleged services NFS : with intensive operations on the attributes
or intensive operations on the data.
Comparison of applications
Each with a different set of NFS In general, applications using the many small
files that can be characterized as performing intensive operations attributes.
Perhaps the best example of this is a classic application of software
development. Large software systems usually consist of thousands of small
modules. Each module contains a file inclusion (include file), source code file,
object files, and some of the archive file type (like SCCS or RCS). Most files
are small, often ranging from 4 to 100 MB. As usual during service transactions
NFS interrogator locks, the processing of these applications is the speed of
light server requests attributes. The total number of transactions over the
occupied less than 40%. In most servers with very intensive operations up to the
attributes required only moderate bandwidth network : bandwidth Ethernet (10
Mbit / s) is usually adequate. Most servers home directory (home directory) fall
into the category of intensive operations to the attributes : most small files
stored. In addition, these files are small compared with the size of the
attributes, they also provide the opportunity to client data file system cache,
eliminating the need for re-enacted from the server. Applications running very
files are categorized intensive operations with the data. This category includes,
for example, the application of geophysics, image processing and electronic CAD.
These annexes usual scenario of NFS workstations or computers includes : reading
very large file, a long process of the file (minutes or even hours), and finally
reverse record a smaller file a result. Files in these applications often reach
of 1 GB and files larger than 200 MB is the rule rather than the exception. When
handling large files dominated operations associated with data requests. For
applications with intensive operations up to the sufficient bandwidth network is
always critical. For example, data transfer speed among Ethernet is 10 Mbps.
This speed seems quite high, but 10 Mbps is only 1.25 MB / s, and even that
speed in practice can not be achieved because of the overhead protocol exchange
and the limited speed of each of the interacting systems. As a result, the real
speed limit Ethernet is approximately 1 MB / s. But even that speed is
achievable only in nearly ideal conditions, when the entire Ethernet bandwidth
to transfer data between the two systems. Unfortunately, the organization has
not usebal, when in fact often the case that only a small number of clients
request data simultaneously. While there are many active clients saturation of
the network is approximately 35%, which corresponds aggregated speed 440
kb / s. The very nature of these types of customers with intensive
implementation of the data determines the planning system configuration. It is
usually the choice network environment, and often dictates the type of server.
In many cases the development of applications with intensive operations with a
need pereprokladki networks. In general, it is believed that in an environment
of the intensive implementation of the data, some more than half of NFS linked
to the transfer of user data. As a representative of intensive operations up to
the attributes of a classic is usually a mixture Legato, in which 22% of all
transactions are read (read) and 15% - operation (write).
A model example of using NFS
After all examples of the use of most applications show that the server
customers burden very unevenly. Consider working with a typical application.
Typically, a user must first take binary code applications that accomplish the
part of the code, which is responsible for organizing the dialogue with the user,
who must determine is required by the data set. The annex reads data set from
the disk (possibly remote). The user interacts with applications manipulating
data in the main memory. This phase has continued much of the time of
application until the end, a modified set of data saved to disk. Most (but not
all) applications that are universal scheme work, often with repetitive phases.
The following figures illustrate the typical load NFS. Figure 4.2 shows a piece
for the magazine SunNetManager 486/33 PC running MS-DOS. The explosive nature of
the clients is very clearly : in short intervals visible peaks reaching as high
as 100 transactions per second, but the average load is small-7 operations per
second, and perhaps typical load is about 1 transaction per second. The schedule
released at intervals measured in one second to see the speed of transactions
with small granularity. Figure 4.3 shows the magazine piece SunNetManager for
diskless client-SPARCstation ELC with a 16 MB memory, performing various
instruments of office automation. Relatively flat load reflected on this chart
is typical of the majority of clients (Lotus 1-2-3, Interleaf 5.3, OpenWindows
DeskSet, email very files). While there are a few cases where speed of 40-50 in
the second, they all have short (1-5 seconds). The average time of the resulting
total load is much lower : in this case, substantially below 1 operations per
second, even without taking into account the free night. The graph measurement
interval of 10 minutes. Note that this is a diskless system with a relatively
small memory. The pressures from clients, with great drive and RAM will be even
less. Finally, Figure 4.4 shows how random nature of the various customer has
the effect of smoothing the load on the server. Chart shows pressure on the
server twenty diskless clients with 16 MB of memory within ten days.
Operating Systems
Operating real memory of a personal computer using a simple two-tier model of I
/ O, in which the main memory and the input / output files are managed
separately. In practice, this leads to even less load on the subsystem I / O.
For example, when the PC under Windows is for Lotus 1-2-3, the entire 123.exe
copied to the main memory system. The main memory is copied to the full code of
1.5 megabytes, even if the user would then assume command quit without
performing any other function. During the execution of the client application
will not issue any additional requests for input / output of the file, since all
binary code is rezidentno memory. Even if this code svopiruetsya Windows, it
will kPa at the local drive, which eliminates network traffic. In contrast, a
system based on Solaris, calling annex copy in memory function quit, and only
those functions that are required to perform initialization. Other functions are
loaded into memory pages later, in actual use, which results in considerable
savings in primary and allocate the time pressures on the subsystem I / O. If
the customer does not have enough memory, the pages can be removed and then
restored to the original source code programming (network server), but it placed
an additional strain on the server. As a result, the workload of the subsystem I
/ O server from PK-klientov is much more cataclysmic than to customers
workstations, with the same application. Another feature of the PC user base is
that the files used by these customers, significantly smaller than the same
files used on the workstations. On the very few PC applications that can be said
is that they are "intensive use of data (see sect. 3.1.3), mainly because the
memory management in the PC operating systems difficult and limited in capacity.
The very nature of the environment associated with intensive work with the
attributes, the choice of configurations to meet the challenges of random access.
While the fastest PC currently on CPU performance may well challenge the
supremacy workstation entry-level model PC is a much less demanding online
customer than the typical workstation. In part, this is because the vast
majority of PCs are still based on a slower processor 386 (or even 286), and
slower processors tend to operate with less demanding applications and users.
Moreover, the slower processors, even working at full speed, simply generate
queries less rapidly than workstations because the internal bus and network
adapters such PCs are not well optimized compared with the devices of greater
size. For example ISA standard Ethernet adapters, available in 1991 were able to
maintain the speed of data transmission only at the level of 700 MB / s (compared
with a rate greater than 1 MB / s, which was achieved in all workstations 1991)
and some fairly common interface cards were only able to provide speeds of
around 400 KB / s. Some PCs, including portable use interfaces, Ethernet, which
are actually connected via the parallel port. While such connections saves slot
bus and conveniently enough, but the Ethernet interface is one of the slow, as
many of the parallel port is limited data transmission speed of 500-800 kbps (60-100
MB / s). Of course when the user base have become a PC based on the 486,
equipped 32-bitovymi network adapters DMA, these distinctions are being blurred,
but is useful to remember that the vast majority of customers PC-NFS (especially
in the country) are in an older, less demanding users. The PC based on the 33
MHz 486DX, 32-bitovym equipped with Ethernet interface, shown in Figure 4.2.
The NFS Client-based UNIX systems such as Solaris the NFS client subsystem
equivalent disk subsystems, namely, it provides a service manager virtual memory,
and in particular the file system on the same basis as disk service, except that
this service is implemented with the assistance of the network. This may seem
obvious, but has some influence on the work of NFS client / server. In
particular, the virtual memory manager is located between client applications
and NFS. Implementing applications requesting the file system are cached system
virtual memory customers, reducing customer requirements for entry / withdrawal.
This can be seen in Figure 4.5. For most applications more memory on the client
to be less load on the server and more common (ie, client / server) system
performance. This is especially true for diskless clients who must use NFS as an
external storage device for anonymous memory. The mechanisms of virtual memory
caching delays, and sometimes cancels work NFS. For example, the diskless
workstation, serving 1-2-3. If data and binary codes applications remotely
located, the system will, as required, to download pages in memory performed
binary codes 1-2-3 with NFS. Then, using NFS to be loaded into memory data. For
most files typically configured 1-2-3 on the workstation will be cache
memory and stay there for a considerable time (more minutes, rather than seconds).
If opens and remains open temporary file, the file is opening itself to both the
client and server, but all updates the file usually are cached for a short time
in front of the client to the server. The semantics UNIX-fayla when file is
closed all changes must be written to external memory device, in this case the
NFS server. Alternatively, cached records may be recorded in the external
storage device using demons fsflush (Solaris 2.x) or udpated (Solaris 1.x). As
with conventional disk I / O, cached data input / output NFS remain in memory
until the memory is not needed for any other purpose. When a list issued to the
server, it must put the data in a stable memory before submission. However, the
client is a little different. If the user goes back to the cached data, for
example, if our example again handled some text pages 1-2-3, instead of issuing
queries to the server, treatment granted directly from the client's virtual
memory. Of course when the customer does not have enough memory to make room for
new data quickly modified pages are written back to the server, and unmodified
pages simply excluded. Since Sun Solaris 2.3 offers a new opportunity, called a
file system with data replication and caching file system (CFS - Cashed File
System). In accordance with standard protocol NFS files selected block by block
directly from the server to your client and the manipulation of them happened in
the memory. Data written back to disk. Software CFS is located between code NFS
client and server NFS access methods. When blocks of data obtained NFS client
software, they are cached in the highlighted area on the local hard disk. Local
copy of the file called the front Plan (front file), and a copy of the file-server
back (back file). Any subsequent appeal to cached file is copied to the
local disk, rather than copies, located on the server. For obvious reasons, such
an organization can significantly reduce the load on the server. Unfortunately,
CFS is not a comprehensive means to reduce the burden on the server. First,
because it did provide copies of the data, the system must provide certain
activities to maintain a coherent state of copies. In particular, CFS subsystem
periodically checks the file attributes back (frequency of testing a user). If
the file back plan has been modified, the front plan vychischaetsya file from
the cache, and the subsequent appeal to the (logical) file will mean that he
will be re-selected from the server and cached. Unfortunately, most applications
continue to work with a file, rather than certain blocks of data. For example,
vi, 1-2-3 and ProEngineer read and write data files its entirety, regardless of
the actual purpose of the user. (Generally, programs that use to access files
command mmap (2) do not apply to the file as a whole, while programs that use
the command read (2) and write (2) generally do). As a result, CFS usually
caches entire file. The NFS file systems subjected to frequent changes are not
very good candidates for CFS : files will be permanently cache and
cleaned, which eventually leads to an overall increase in network traffic,
compared to the simple work through NFS. The challenge of maintaining a coherent
state of the cached data between client and server is also another problem :
when a customer modifies the file, the file forward plan is cancelled, and file
back plan appropriately updated. Subsequent treatment reading of the file will
choose and then cache file. If updating files is a standard practice, it leads
to more traffic than when the standard NFS. The CSF is a relatively new
opportunity, unfortunately very little has been done measuring its conduct in
actual use. However, the very idea of a protocol CSF leads to the following
recommendations : The CSF should be used for the file system, which is mainly
read data, such as shared file systems coding applications. CSF is
particularly useful for sharing data between relatively slow network, such as
WAN, connected by lines less than T1. CSF useful for high-speed networks,
interconnected routers, which make the delay.
Configuring NFS-server
Assumptions To collect sufficient and accurate information to create NFS server
configuration should answer the following questions : Is the intense pressure on
the attributes or intensive data? Will customers to use a caching file system
to reduce demand? How many average will be fully active clients? What types
of client systems to be used and under what operating systems they work? How
big file systems to be used in the separation of access? Repeat if demands of
different customers to the same file (for example, to include files), or they
belong to different files? What are the number and type of alleged exploitation
networks? Is the current network configuration suitable for the type of traffic?
Is the purported server configuration of the CPU to manage traffic associated
with the applicable network? If local network (WAN), whether the medium and
routers rather small delay and higher bandwidth to ensure the practicality of
NFS? Is the disk drives and SCSI host adapters to the main set of productivity?
Is the use of software tools like Online : DiskSuit to adequately burden on
access to records of all the disk drives? If used frequently writes NFS, does
the configuration of NVSIMM? Does the proposed strategy backup type, number
and location of the SCSI bus devices backup? Perhaps the most important
requirement for NFS-servera configuration is sufficient bandwidth and
preparedness network. This practice leads to the need for a configuration with
the number and type of networks and interfaces. As noted earlier, the most
important factor in determining the choice of network configuration, is the
dominant type of NFS used applications. For applications with intense pressure
on the required relatively few networks, but those networks must have high
bandwidth, such as FDDI network or CDDI. These requirements may be met with a
100baseT networks (Ethernet 100 Mbit / s) or ATM (Asynchronous Transfer Mode 155
Mbps). Most of the attributes intensive applications and working with a less
expensive infrastructure, but may require a large number of networks. To decide
on the choice of network is relatively easy. If for individual customer needs
aggregated speed in excess of 1 MB / s, or for simultaneous operation of
multiple customers need network bandwidth in excess of 1 MB / s, these
applications require high-speed networks. There are in fact (1 Mb / s)
artificially inflated, because it describes the speed that you do not exceed the
guarantee. Usually considered more prudent speed Ethernet network of
approximately 440 MB / s, but not limited to 1 MB / s. (Typically, users
perceive Ethernet as neotvechayuschuyu "is about 35% in network load. The figure
is 440 MB / s corresponds to 35% tax load lines with a capacity of 1.25 MB / s).
If you do the customary mode does not require bandwidth, it may be less than
adequate speed network environment like Ethernet or TokenRing. This environment
provides sufficient speed in the performance of lookup and getattr that dominate
annex intensive attributes and the relatively easy data traffic associated with
such use. High-speed networks are most useful for large groups of customers with
intense pressure on the rather because of the lower cost of infrastructure,
rather than for reasons of maximum capacity, in collaboration with one of the
other. This is the current state of NFS protocol, which is currently working
with blocks of length 8 MB and provides pre-selected only 8 MB (ie one operation
to the server can be defined as 16 MB of data). The overall effect of such an
organization is that the maximum data transfer speed between the client and
server to interact via FDDI ring is approximately 2.7 MB / s. (This speed is
achieved by using the file / etc / system on the customer operator set nfs :
nfs_async_threads = 16. Customers must run SunOS 4.1.x 12 biod demons, and not
eight as it is by default). The speed of three times the maximum speed, which
provides Ethernet though the speed of FDDI ten times greater. (NFS is an
application layer protocol (level 7 in the OSI model). Minutes lower levels,
such as TCP and UDP can handle much higher speeds, using the same hardware. Most
of the time waiting for responses and other processing application layer. Other
application layer protocols, not designed for immediate response and / or
confirmation, also can convey a much higher rate environment). Peak speed in the
case of 16 Mbps Token Ring is approximately 1.4 MB / s. More recently announced
a new version of NFS protocol +, which eliminates this limitation, allowing work
with much larger units. NFS + allows transmission blocks almost arbitrary size.
The client and server agree on the maximum amount for each unit to mount a file
system. The block size can grow up to 4 GB. The main advantage 100-Mbitnyh
networks using conventional versions of NFS is that the network can support many
simultaneous transmission of data without degradation. When the server sends the
data to Ethernet clients at the rate of 1 MB / s, the transmission consumes 100%
available network bandwidth. Attempts to transfer to the network more data lead
to a lower capacity for all users. The same client and the server can send data
at a speed of 2.7 MB / sec for FDDI rings, but in a high-speed network that
transaction consumes only 21% of available bandwidth. The network can support
five or six channel at the same time without serious degradation. This situation
can be likened to a high-speed backbone. When the traffic light (light traffic)
speed highway with two lanes and a speed limit of 90 kilometres per hour is
almost as good as vosmipolosnaya superhighway with a speed limit of 120 km per
hour. But when traffic is heavy (heavy traffic) Mainline much less susceptible
to congestion. The FDDI also slightly (about 5%) is more effective than Ethernet
and Token Ring environments in intensive sending data because it can put the
package more useful data (4500 bytes, compared to 1500 bytes to 2048 bytes
Ethernet and Token Ring to). When sending data of 8 MB needs to process only two
sets, compared with five or six for Token Ring and Ethernet. But all these
considerations are valid only for the environment with intensive data transfer,
as the attributes in the processing of requests is small (for 80-128 bytes), to
transfer only one package regardless of the type of network in use. If the
existing wiring in the enterprise network pre-empt the use of FDDI fiber, there
are standards for copper wire FDDI (CDDI), which ensures that switching network
with the existing facility on the basis of TP cable. While ATM has not yet
become a widely used technology, perhaps in the future it will become the main
vehicle for protection from intense sending data because it provides faster data
(currently defined data transmission speed 155 Mbits / s, 622 Mbit / s and 2.4
Gbit / s) and the use of point-to-point topology, in which each joint klient-hab
can work with a speed of its environment.
NFS and global networks.
in real life situations to arise where the client and server NFS may be located
in different networks, integrated routers. Topology network can greatly affect
the perceived user productivity and NFS server provided by the service. The
effectiveness of NFS-servisa through integrated network should be carefully
monitored. But at least know that you can successfully configure the network and
applications in the global (wide) topology NFS. Perhaps the most important issue
in this situation is the delay of operation : the time that elapses between
receipt of the extradition request and response. The delay in the implementation
of LAN not so much because of such networks relatively short distances can cause
significant delays in data transmission environments. The global network delay
operations can occur only when carrying packets from one point to another. Delay
packet consists of several components : Delay router : routers spend a long (and
often substantial) time to do proper routing packets from one network to another.
Note that in constructing the majority of global networks (even in the
construction of the line between two adjacent buildings), at least two router.
Figure 4.6 provides topology typical university campus, which is usually between
client and server, three or even four router. Delayed transmission network :
physical medium used to transmit packets through the global network, can often
make their own significant delay beyond the largest delay routers. For example,
satellite bridges often associated with the emergence of a very long delays.
Faulty transmission : a global network for possible order of magnitude more
sensitive to errors transfer than most local area networks. These errors cause
significant re-channel data, resulting in increased delays for operations, and
to reduce the effective capacity of the network. The network is highly prone to
error transfer of a block of data NFS often set equal to 1 MB instead of the
normal 8 MB. This reduces the amount of data that must be rebroadcast in the
event of errors. If an acceptable level of a litmus transfers data file service
on the global network is possible. Most of the configuration of the network used
high-speed synchronous serial point-to-point links, which are connected to one
or more local area networks at each end. In the United States, such successive
lines usually speed 1.544 Mbps (T1 line), or 56 Kbps. European communication
companies offer a little more speed : 2.048 Mbps (E1 lines) or 64 kbit / s
respectively. There are even more high-speed data link. These leased lines,
known as T3, offer transmission speeds up to 45 Mbps (5.3 MB / sec). Today, the
majority of T3 lines used for data transmission. At first glance, it seems that
this line significantly slower than local networks to which they are connected.
However, the rapid sequential line (T1) provide bandwidth is much closer to the
real capacity of local networks. This is because successive lines can be used
with almost 100% of capacity without incurring excessive overhead costs, while
Ethernet usually hurt some already with 440 MB / s (3.5 Mbit / s), which is only
about twice the bandwidth T1 line. For this reason, file service for high-speed
serial links is possible and can transfer data at an acceptable speed. In
particular, the organization is useful when transferring data between remote
offices. The annexes to the intensive processing attributes NFS work on global
network can be successful if the delay operations is critical. The global
network of short packets are transmitted through each segment promptly (with
high capacity), although delays routing and the environment often cause
considerable delays in the operation. Conclusions : To the global services NFS
suited consistent line T1, E1, or T3. For most of the NFS line at speeds of 56
and 64 kbit / s are usually not been fast enough. When NFS through the global
network there are problems with delays and network routing. The capacity of the
network is usually no problem. To significantly reduce traffic on the global
network, you can use a client of caching file system (CFS), unless this traffic
is dominated writes NFS. Given these considerations, to determine the proper
type and number of networks can be used following empirical rule : If annexed
dominated operations on the data, choose FDDI network or any other high-speed
network. If, for logistical reasons, the laying of fibre-optic cabling is not
possible, consideration should be given to implementing FDDI on twisted pair.
When a new system should be borne in mind that ATM networks used the same cables
as for FDDI. The network configuration should be one FDDI ring for each customer
5-7, while fully active in the sense of NFS and intensively working with the
data. It should be remembered that very few intensive data applications
continuously generate queries to the server NFS. In typical intensive data
applications automate the design of electronic devices and systems studies the
Earth's resources are often allowed to have up to 25-40 clients for the ring. In
systems with intensive data, where the existing system is forcing cable to use
Ethernet, a separate Ethernet network for every two active clients and customers
as 4-6 to the same network. If the annex involves intensive processing
attributes, it is enough to build Ethernet or Token Ring. In an environment
with a high use of attributes should have one Ethernet network to 8-10 all
active clients. Close exceed 20-25 clients on Ethernet regardless of
expectations because of the sharp degradation caused activity in the case of
many customers. As a reference point in terms of common sense suggests that
Ethernet can support 250-300 NFS-operatsy per second to test SPECsfs_97 (LADDIS)
even with a high level of conflict. Nor exceed 200 NFS operations per second
state security. configure a network for every 10-15 TokenRing fully active
clients in an environment with a high use attributes. If necessary, the Token
Ring network can be connected with an excellent 50-80 clients characteristics of
this type of network on the resistance to degradation under heavy load (over
Ethernet). For systems that provide multiple classes of service users make
sense of mixed network configuration. For example, and FDDI and Token Ring
suited for the server, which supports both applications related to the display
of (intensive data), and PC group performing annex financial analysis (perhaps
intensified by attributes). Because many computers are universal system, which
allow a great increase in the number connected to peripheral devices, almost
always possible to configure the system so that the main limiting factor is the
processor. Under NFS power processor is spent directly for processing protocols
IP, UDP, RPC and NFS, as well as for managing devices (disk and network adapters)
and the manipulation of the file system (blatantly suggests that the consumption
of CPU time is growing proportionately in accordance with the order here). For
example, Sun recommends the following empirical rules for configuring NFS-serverov
: If the customer is dominated by intense attributes Wednesday and is less than
4-6 Ethernet or Token Ring, then to work as NFS-servera enough odnoprotsessornoy
system. For systems with one or two smaller networks enough processor power
machines entry-level SPARCserver 4. For very intense on the attributes of many
network recommended dual system such SPARCstation 20 Model 502 or dual
configuration SPARCserver SPARCcenter 1000 or 2000. If Wednesday intensive data,
it is recommended to configure two SuperSPARC processor with a high-speed
SuperCashe per network (such FDDI). If the existing restrictions on the
organization wire dictates the use of Ethernet in such an environment, it is
recommended to configure a processor SuperSPARC for every 4 Ethernet or Token
Ring.
Configuration of disk subsystems and load balancing
configuration like network configuration is determined by the type of client
disks. Productivity disk drives varies widely depending on the requirements of
their access methods. The random access nature almost always nekeshiruem and
demands that were actually let positioning disk for each I / O (mechanical
movement, which significantly reduces productivity). When serial access,
particularly consecutive hits in reading, much less mechanical movement caret
drive to the operation (usually one per cylinder, approximately 1 MB), which
gives a much higher capacity. Experience shows that the majority of applications
to run in an environment with intensive data is consistent, even on the servers
that provide data to many clients. However, as a rule, the operating system has
been working hard to provide access to their devices. So if you want to provide
services for applications with intensive data should be chosen configuration for
a coherent environment. For example, his drive capacity 2.9 GB disk was the
fastest for Sun consecutive applications. He could provide data via the file
system at speeds of 4.25 MB / sec. It was also the most concise and drive Sun
has been the most convenient for storing large amounts of data. High speed data
to the SCSI bus speed (peak bus bandwidth is 20 Mbytes / s) determines the
optimal configuration of the disk subsystem : 4-5 active disk capacity 2.9 GB on
one main adapter (DWI / S). If additional capacity to store data, the inclusion
of more disks for each primary adapter is perfectly acceptable, but it will not
increase the performance of disk subsystems. Discs with 2.9 gigabytes of Sun
posted on the set in a rack chassis (up to six disk drives to the chassis). Each
chassis can be connected to two independent SCSI adapters principal. This
possibility is recommended to configure servers, serving customers with
intensive data requests. To ensure maximum disk storage capacity of up to 12
drives can be configured on a single adapter DWI / S. However, maximum
performance only 4-5 reservoirs. In an environment with consistent access to
calculate just how many disks required for peak periods. Every fully active
client may require the disk subsystem capacity to 2.7 megabytes per second. (This
assumes the use of high-speed transmission networks in a 100 Mbit / s and above).
A good first approximation to a single 2.9 GB disk for every three fully active
clients. It is a ratio, while each drive can transfer data at a rate of more
than 4 megabytes per second, and customers seek only 2.7 MB / s, since the two
active clients on the same disk will be constant movement forward and backward
caret between the cylinder (or file systems), leading to significantly lower
capacity. To balance the work drive, as well as to accelerate certain types of
channel data can use special software like Online : DiskSuit 2.0 (Sec. 4.3.4.3).
If a network environment applies Ethernet or 16 Mbit Token Ring, one disc for
each fully active user. If using NFS + is the attitude changed greatly since NFS
+ provides individual capacity in the client / server about speed network
environment. Unlike environments with heavy data really all accessing files
among intensive attributes to arbitrary access to the discs. When files are
small, access to data dominates sample lines directories lines index descriptors
and the first few blocks of indirect (positioning required to get all the pieces
really meta), and each unit of the user. As a result of a carriage drive spends
much more time "ryskaya" between different pieces of meta file system than the
actual sample of the user. As a result, the criteria for the choice of intensive
NFS attributes to be materially different from the criteria for the environment,
with intensive use of the data. As in the time required for the operation of
random I / O dominated time positioning caret drive, the total capacity of the
disk in this mode is much less than in the sequential access. For example, the
standard disk drive production in 1993 to work at a speed of 3.5-4 MB / s in
sequential mode access, but provides the only access arbitrary 60-72 operations
per second, or about the speed of 500 kb / s. In these circumstances, SCSI bus
is much less taken, which allows the user to drive it much more before it came
to handling tires. In addition, one of the tasks of configuring the system is
the most reasonable number of disk drives, as it determines the number of disc
carriage, which is a limiting factor in disk subsystems. Fortunately, the very
nature of the attributes intensive applications suggests that the requirements
of disk storage is relatively small (relative to the intensive data applications).
In these circumstances often useful to include in the system configuration,
instead of one large drive two or even four smaller capacity drive. While this
configuration is slightly more expensive in cost per megabyte of memory, its
performance has improved. For instance, two 1.05-GB disk capacity is about 15%
more expensive than one disk capacity of 2.1 GB, but they provide more than
twice the capacity of random I / O. Around the same attitude is just between
drive capacity of 535 MB and 1.05 GB drive (see Table 4.2). Thus, for intensive
on the attributes of a better environment to configure many small disks
connected to the main moderate SCSI adapters. Disk capacity of 1.05 GB has
excellent proprietary software, which minimizes the load SCSI bus. Disk capacity
of 535 MB is similar characteristics. Recommended configuration is fully active
4-5 535 MB or 1 GB drives per SCSI bus, while six or seven disks can work
without causing serious conflicts on the bus. The very system intensive features
that require the use of disk capacity of 2.9 GB (for reasons of server or data),
the best performance in 8 fully active disk on the bus fast / wide SCSI but can
be fitted and 9 or 10 disk drives with only a small degradation of response time,
I / O. As with intensive data systems configuration more drives on the SCSI bus
provides additional memory capacity, but did not give further progress in
productivity. Difficult to give any specific recommendations on the number of
disc carriage, which requires intensive attributes on Wednesday, as the load
varies widely. In such an environment, server response time depends on how
quickly attributes can be found and handed over to the client. Experience has
shown that often useful to configure at least one disk drive for every two fully
active clients to minimize delays do these things, but the delay can be reduced
with the help of additional main memory, allowing cache frequently used
attributes. For this reason, lower-capacity drives are often preferred. For
example, better use eight drive capacity of 535 MB of disk instead of two 2.1 GB
capacity. One of the common problems is the lack of NFS server load balancing
between disk drives and disk controllers. For example, to support a number of
diskless clients often used configuration of the three disks. The first disc
contains the operating system and server applications binary codes; Second-disk
file system root and swap for all the diskless clients, and the third drive-user
home directories diskless clients. This configuration balanced as possible on a
logical rather than a real physical pressure on the disc. In such an environment,
drive, store swap area (swap) for diskless clients are usually much more so than
any of the other two discs : the drive almost all the time will be loaded to 100%
and the other two by an average of 5%. Such distribution is often used also to
work in another environment. To ensure transparent distribution of access to
multiple disk drives can be used successfully as fission and / or mirroring,
supported by a software type Online : DiskSuit. (Konkatenatsiya drive is minimal
load balancing, but only when the drives are relatively well). In an environment
with a high use of fission with a small cut provides increased capacity drives,
as well as distribution services. Splitting drive significantly improves the
performance of a consistent reading and writing. A good initial approach to
value is the ratio of floors 64 MB / number of disks in the strip. In an
environment with a high use of attributes that are characterized by random
access, the default floors (one disc cylinder) is the most appropriate. While
disk mirroring feature in DiskSuit primarily designed to ensure the
sustainability of the prevention, the side effect of its use is to improve the
access time and reduce the burden on disks by providing access to two or three
copies of the same data. This is especially true for the environment, which is
dominated read. Operation record on mirrored disk usually performed more slowly,
as each sought logical operations actually required two or three operation.
Currently, the computer industry recommended maximum load of each disk drive in
the 60-65%. (The boot disk can find using the iostat (1)). Usually, in practice,
not pre-planned distribution of data so as to ensure that the recommended boot
drive. This required several iterations, which include removal of reorganization
and related data. Moreover, it should be noted that the standard distribution of
CDs with the times changing, sometimes dramatically. The distribution of the
data, which provided very good job at the time of installation, may be very weak
results a year later. When optimizing data on the existing set of disk drives
are many other considerations second order.
The best areas disk drives.
Many are now supplied computer companies, benefit from the code, which are
called "zone-bit recording zone bit recording). This type of encryption allows
the use of the geometric properties of a rotating disk pack more data on the
surface of the disk, which are further away from its center. The practical
effect is that the number of addresses with lower numbers (corresponding to the
external disk cylinders) exceeded the number of addresses with large numbers.
Usually this limit is 50%. This type of coding increasingly affects productivity
consistent access to the data (for example, 2.1 GB disk indicates data
transmission speed range 3.5-5.1 MB / s), but it also affects the performance of
random access to the drive. Data located on the outer cylinder drive, not only
are faster under head read / write (and therefore speed above), but these
cylinders as simply more in size. Assigned amount of data can be divided into a
smaller number of large cylinders, leading to fewer mechanical movements caret.
Ground rules for configuring the drive can be summarized as follows :-In an
environment with intensive data should be configured from 3 to 5 fully active
2.9 GB disks for each primary adapter fast / wide SCSI. Must include at least
three disk drive for each active client with FDDI network, or a CD for each
active client with Ethernet or Token Ring. In an environment with a high use
of attributes should be configured around 4-5 fully active 1.05 GB or 535 MB of
disk for each primary SCSI adapter. Must include at least one CD for every two
fully active clients (in any network environment). Each primary adapter can
connect additional drives without substantial degradation of performance, until
the number of normally active drives on the SCSI bus does not exceed the
guidance given above. To load access multiple disks can recommend the use of
software such as Online : DiskSuit 2.0. If possible, use the fastest zone on
the disk. As many UNIX-sistemah (eg, Solaris) implemented caching files in
virtual memory, most users tend to configure servers NFS very large main memory
subsystems. But examples of typical use NFS client files show that the data
retrieved from the buffer cache in fact relatively rare, and additional main
memory is usually not necessary. In typical server space allowed customers to
drive far exceeds the amount of main memory system. Repeated requests for blocks
of data rather satisfied by reading from memory rather than from disk.
Unfortunately, most clients work with their own files and rarely used common
shared files. Moreover, the majority of applications are usually read the entire
file into memory, then closes it. The client rarely drawn to the original file
again. If none of the other customers did not use the same file before
overwriting data in a cache, it means that the cached data should never again be
used. However no additional overhead costs associated with caching of data, and
their subsequent failure to use available. If you need memory to cache the other
pages of data, in the pages just happens to record new data. Not necessarily
dump page to disk, since memory to be retained is already on the disk. Certainly
little of storage in the memory of the old pages, which are not used. When the
amount of free memory falls below 1 / 16 of total memory pages unused in the
recent past, become available for reuse. The biggest exception to this rule is
empirical temporary file, which often opens early and closes at the end of the
work. Since the file remains open for the client, the data associated with the
file, be displayed (and cached) in the memory of the client. Virtual memory
subsystem uses a client server as a backup storage temporary file. If the
customer does not have enough physical memory to store these pages in their
cache, some or all of the data will otkacheny or replaced by new content in
meeting future operations, and repeated applications of the data will lead to
the extradition read NFS to restore the data. The ability to cache server data
is a certain advantage. Operation record can take advantage of caching UNIX
server, because NFS protocol requires that any transaction records to a remote
disk fully implemented simultaneously, to ensure a consistent state even in the
event of server failure. The operation is called synchronous recording, in that
sense, because the logical operation of the drive in a fully completed before it
is confirmed. Although data queries, writes NFS not benefit from the buffering
and postponing the date, normally performed by a write UNIX. Note that the
client can and usually writes in the cache in the same manner as any other disk
I / O. Customer, in any case, to check the results writes dumped on "disk",
whether local or remote disk. Perhaps the simplest and most useful is the
empirical rule, which is called the "rule of five minutes." This rule is widely
used in the configuration database servers, much greater complexity which makes
it very difficult to determine the size of the cache. The current ratio between
the prices of memory and disk subsystem shows that the cost-effectiveness of
data cache, which made treatment more than once every five minutes. In
conclusion, a few simple empirical rules, which allow you to choose the
configuration memory of NFS servers : If the server essentially provides
custom data many customers to configure on the minimum memory. For small teams
usually its volume is 32 megabytes, and for large groups of about 128 megabytes.
The SMP configuration, always have at least 64 MB for each processor.
Applications with intensive use of the attributes usually benefit from increased
system memory slightly higher than applications with intensive data. If the
allocated space on the server to store temporary files that will work very
intensively with these files (a good example is the company Cadence Verilog)
should be configured server memory roughly equal amount of active temporary
files used on the server. For example, if the temporary file client about 5 MB
and it is anticipated that the server will serve 20 fully active clients, there
should be a (20 customers x 5 MB) = 100 MB of memory (128 MB of course is the
most convenient target, which is easily configured). Often, however, this
temporary file can be placed in a local directory such as / tmp, resulting in
significantly higher performance customer, as well as significantly reduce
network traffic. If the main task of the server is storing only executable files
should be configured server memory about equal stock of heavily used binary. Do
not forget the library! For example, the server predpolagayuschiysya storage /
usr / openwin for some team members must have sufficient memory to cache Xsun,
cmdtool, libX11.so, and libxwiew.so libXt.so. This application NFS differs
significantly from a server model so that it is supplying the same files to all
its customers and therefore is able to cache the data. Typically, customers do
not use every page of all binaries, and so wise in the server configuration to
provide only the amount of memory, enough to accommodate frequently used
programs and libraries. The memory can be selected based on the five minute
rule : 16 MB for the operating system plus memory for caching data, which will
occur treatment more often than once every five minutes. Because NFS servers
fail to meet user processes, a large swap space (swap) is not needed. For
servers swap space of approximately 50% of the size of main memory is more than
enough. Note that the majority of the exact opposite of what everyone expects.
Disc operations inherently associated with mechanical movements heads in the
drive, so they are slow to be fulfilled. Usually UNIX buffers writes to main
memory and allows the issuing process continues, while the operating system will
have to physically write data on the disk. Simultaneous principle writes in NFS
means that they normally very slow (much slower than writes to a local drive).
If the client issues a request records, requires that the server has updated the
data on disk, and all associated metadata file system. For a typical file should
be completed by 4 disk : each operation must update the data information in the
directory file inditsiruyuschuyu date of last modification, and indirect heat;
If the file is large, it will also update the second indirect heat. Before
confirm that the request for records NFS, the server must implement all these
updates and to ensure that they were on the disk. Operation record NFS can often
last for 150-200 milliseconds (three or four simultaneous recording more than 40
milliseconds each), compared with the normal 15-20 milliseconds to write to your
local disk. In order to accelerate the writing NFS, servers can use the stable
memory (RAM-Non-NVRAM). This additional opportunity is based on the fact that
the NFS protocol simply requires that the data writes NFS would be recorded in
stable memory instead of fixing them on the disk. Until the server returns data,
which confirmed the previous transaction records, it can keep the data in every
way possible. PrestoServe and NVRAM exactly turning this semantics. The
installation of these devices to the server device driver NVRAM intercepts
requests synchronous writes to disk. These are not sent directly to the drive.
Instead, the write down in a stable memory and confirmed as completed. This is
much faster than waiting for the end of the mechanical operation of data on the
disk. A short time later, the data are recorded on the disk. As a logical
transaction records NFS performs three or four synchronous disk operations, the
use of NVRAM significantly improves bandwidth writes NFS. Depending on the terms
(of the file system, there are other requests to the disk, the size and location
of records, etc.) using NVRAM accelerates NFS writes in 2-4 times. For example,
a typical capacity in the discharge writes NFS running Solaris 2 is about 450 MB
/ s. Using NVRAM speed increased to about 950 MB / s and even higher if the
network environment faster than Ethernet. There is no better time of the read
NVRAM not. From the point of view of the disk subsystem or NFS clients
additional opportunities PrestoServe and NVSIMM functionally equivalent. The
main difference is that NVSIMM more effective because they require less
manipulation of the data. The fee PrestoServe physically located on the bus SBus
requires that data be copied to it through peripheral bus. In contrast, NVSIMM
placed directly into the main memory. Recordable Disc data is not copied to
NVSIMM via peripheral bus. Such copying can be done very quickly with the help
of steering. For these reasons NVSIMM are preferable to a situation where both
opportunities and NVSIMM PrestoServe are available. Due to the importance of
accelerating Sun received recommends the use of NVRAM indeed, in all of which
provide universal service NFS. The only exception to this rule are the servers
that provide service only in reading files. The most typical example of such use
are servers, store binary codes for a large group of customers. (The Sun known
as a server / usr / dist or softdist). Because the device driver NVSIMM /
PrestoServe be located on the disk in the root file system, accelerated by NVRAM
can be obtained for the work with the root filesystem. NVRAM driver should be
able to pump modified buffer to disk before any other will be an active process.
If a root file system has been accelerated, it could prove to be "dirty" (retrofit)
after the collapse of the system, NVRAM, and the driver could not boot. Another
important consideration when comparing servers, which are equipped with NVRAM
and without it, is that the use of such acceleration generally reduces the
maximum capacity of about 10%. (Systems using NVRAM, NVRAM cache must manage and
maintain a coherent position copies in the cache and disk space). However, the
response time of an improved (around 40%). For example, the maximum capacity
SPARCserver 1000 to test LADDIS without NVSIMM of 2108 operations per second
with 49.4 ms response time. Studies NVSIMM system can perform only about 1,928
transactions per second, but the average response time is reduced to about 32 ms.
This means that customers perceive the NFS server, equipped NVRAM, much faster
than a server without NVRAM, while the total capacity of slightly reduced.
Fortunately, the value of 10% is rarely a problem, since the maximum capacity of
most systems far exceed the typical load, which is in the range of 10-150 per
second on the network.
The backup and fault-tolerance problems.
backup copy file systems and sustainability for failure to NFS server similar to
the problems encountered in the operation of any other system. Some suggestions
for backup and fault tolerance can be summarized as follows : A simple,
relatively small backups can be made by one or two tape drives. The location of
these drives on the SCSI bus does not make much difference if they are not
active during working hours system. Establishment of full backups requires
locking file system to prevent modification. For such operations require special
software, such as product Online : Backup 2.0. As in the previous case, the
location of backup SCSI bus does not make much difference if the copying is
performed outside working hours. Zerkalirovannye file systems enable face full
waivers drive and, in addition, allow continuous access to even during a fully
coherent backups. Mirroring results in a very small loss of capacity disks to
write operation (up to 7-8% at random, and by 15-20% in the sequential access;
In a system with many users, which includes the majority of servers, it is hoped
that those numbers will decrease by half). Mirroring automatically improves
capacity in the discharge of reading. The establishment zerkalirovannyh file
systems, each mirror must be configured on a single SCSI bus. If backups should
be done during normal working hours system, the device must be configured to
copy or on its own SCSI bus, or on the same SCSI bus, which switched from the
mirror (a separate and inactive), to avoid a number of problems with target
response time. When the quick restoration of the file system within the
intensive attributes should be in the configuration NVRAM. In an environment
with intensive data should explore the possibility of high-speed mechanical
devices such stekkerov tapes and storage devices. Work on the assessment burden
on the future system is not very accurate, but often it is a good approach,
which you can obtain in advance. The two basic approaches. The preferred method
is to measure the parameters of the existing system. This method provides some
confidence in the accuracy of the estimate of the burden, at least for the
current time, although not of course guarantee that the pressure in the system
in the future will be equivalent to the existing one. An alternative method is a
rough calculation. It is useful, when the user does not have the necessary
measuring system. To create a sufficiently precise configuration of the system
need to know two things : a mixture of NFS and total system capacity. Mixture of
NFS to show whether the system is intense on the attributes or data. There are
many different mechanisms for the measurement of existing systems. The simplest
of these is simply to use the nfsstat (8), which provides information on the mix
of operations. As these statistics can be re-established through zero-z flag,
the team nfsstat can also be used for the measurement bandwidth using Shell
script, such as shown below. Out of NFS-vyzovov shows, which were served in a
given interval, and therefore the speed with which transactions are processed
NFS. It should be noted that when heavy loads team sleep may actually "sleep"
considerably more than requested by 10 seconds, leading to inaccurate data (ie
re quantity). In these cases must be a better way. There are many such funds, as
described SunNetManager, NetMetrix of Metrix and SharpShooter from AIM
Technologies. These funds make it possible to verify the capacity of the system
under real pressure and mixture operations. To calculate the average capacity
usually requires some further processing of data. You can use a variety of means
(awk (1), spreadsheet type WingZ or 1-2-3). If the measuring current system is
not available, it is often possible rough estimate based on the intended use of
the system. Implementation of the assessment requires an understanding of how
data will be manipulated by the customer. This method is accurate, if it falls
into the category of intensive data. Some judicious assessment usually can be
made and environment-intensive features, but a number of factors make such an
assessment somewhat less accurate. The first step for such an assessment is to
identify fully active query model client. This requires an understanding of
customer behavior. If the intense pressure on these, it makes more sense just
prosummirovat anticipated operations reading and writing and take a number as a
burden for each client. Operation attributes are generally insensitive to the
workload, which is dominated by operations on the data (on the one hand, they
represent only a small percentage of all transactions, on the other hand, these
operations are the minimum number of servers in relation to the work to be
performed for a sample data). For example, the client workstation, performing
software that searches of a given temperature in a certain amount of fluid.
Model data set for this task is 400 megabytes. Usually he read portions of 50
Mbytes. Every piece is fully processed before moving to annex next. The
processing of each segment takes approximately five minutes TSP, and the
resulting files are saved to disk, the size of about 1 MB. Suppose that in a
networked environment using FDDI. Maximum load NFS will arise when the client
reads a portion of 50 Mbytes. The maximum speed of 2.5 MB / s client will be
fully active for about twenty seconds, serving 320 reads per second. Since the
launch of the programme took about 40 minutes (or 2,400 seconds) time, and one
run needed (400 + 1) x 125 Mb ops / Mb = 50,125 ops, the average speed is about
20 ops / sec. The server will provide services to a peak speed of requests (320
ops / sec) for about 20 seconds out of every five minutes, or for about 7% of
the time. This exercise can be drawn three portions of useful information :
average speed active enquiries (20 ops / sec), the peak rate requests (320 ops /
sec) and the likelihood that the peak speed is required. Based on this
information can be formed assessment of the overall rate requests. If the system
configuration will be 10 customers, the average speed of 200 requests ops / sec.
(This speed should not be compared with test results LADDIS because of the
mixture of very different). The likelihood that the client will seek two with a
peak speed of about 0.07 x 0.07 = 0.049, or about 5%, and 3 client will require
peak service only within 0.034% of the time. Thus, from this information prudent
to withdraw the following conclusions : The possibility that the three
customer will be active simultaneously, much less than 1%, the maximum load will
exceed the individual peak load of 2-3. Requires only one network, as the
maximum anticipated load is only 3 x 2.5 Mb / sec = 7.5 MB / s, or far below the
maximum bandwidth of FDDI (12.5 MB / sec). As at any time be fully active only
two or three clients, it takes at least 3 to 6 disk drives (although the model
file size of 400 MB is very likely that it will take more than just six discs
for data storage). At least two major SCSI adapter. Because the system is a
high-speed network, the use of a server with two CPUs SuperSPARC / SuperCashe.
Since it is unlikely that a very large cache file will be useful for a server
requires a minimum amount of main memory, 128 MB is sufficient. If a
relatively small farm drive, for example, of about 16 GB, the system
SPARCstation 10 Model 512 very well be able to cope with this task, as one SBus
slot is required for FDDI interface, and the remaining three slots can be used
to install the main SCSI adapters, to a total of 4 interface FSBE / S, each of
which connects disk drives for a total capacity of 4.2 GB. However, this
application might better suit the system SPARCserver 1000, which will offer a
greater capacity memory : a system with two systemic card allows configuration
with the seven principal adapters and SCSI disk storage capacity of 28 GB (one
mnogodiskovomu device 4.2 GB capacity on each card FSBE / S, not counting the
four fixed disk capacity of 535 MB). In case you need large capacity drives can
be configured system SPARCcenter 2000 with two system boards to give six
interfaces DWI / S up to 12 chassis-drive capacity of 2.9 GB - about 208 GB of
memory. All proposed system can be installed without the use NVSIMM SBus slots,
and all are easy to maintain a required two processors. Using NVSIMM not very
important because the proportion of entries too small (less than 1:400, or 0.25%).
Note that the choice of system configurations for applications with intensive
data generally not very useful to compare the estimated speed rating of requests
from the server to SPECsfs_097, as a mixture of different so that the load can
not be compared. Fortunately, this is usually accurate.
Evaluation of intensive attributes
In the previous example, it was assumed that load NFS operations with the
attributes was negligibly small in comparison with transaction data. If it was
not, for example, in an environment of software development, need to make some
assumptions about the expected mixture NFS commands. In the absence of other
information, the sample can be taken, for example, the so-called mixture Legato.
There SPECsfs_097 (also known as LADDIS) used it is this mixture, in which
operations on the data include 22% of read and write 15%. Consider a client
workstation, the most intensive work which relates to recompile software system
consisting of a source of 25 Mbytes. We know that workstations can compile the
system in about 30 minutes. The compilation generates about 18 megabytes
intermediate object code and binary codes. From this information we can conclude
that the system will save the client to the server 18 Mbytes and read at least
25 MB (maybe more, since almost one-third of the source code consists of header
files are included by multiple source modules). To prevent re-read these files
can be used include a caching file system. Suppose that used CFS. During the "construction
" must allocate 33 MB of real data, or 33 Mb x 125 ops / = 4,125 Mb of data for
30 minutes (1800 seconds), which is approximately 2.3 speed ops / sec. (This
assumes that each operation is performed with 8 Kb of data, so to send 1 Mb of
data required 125 operations). Because it is linked to the intensive use of the
attributes necessary to evaluate a significant number of promahivayuschihsya
attributes. Assuming that the mixture with a mixture of Legato, the overall rate
will be approximately the same : For we also have (18 Mb to record 125 x ops /
Mb) / 15% / 1800 seconds, or 8.33 ops / sec. The combination of reading and
writing very similar to a mixture of Legato, but that might not be the case, for
example, if the files were opened browser source code (the browser files of
source code (browser source files) often 4-6 times the size of the source code).
In this case, we have no way of measuring peak. If there are 20 work stations,
working in the mode described above, we can conclude the following : Even if
it is incredible condition, when all 20 workstations fully active at all times,
the overall rate request is 8.33 ops / sec x 20 clients, or 166 ops / sec, or
below the maximum of 200 ops / sec, which supports Ethernet. Cautious
skonfiguriruyut people for a load of two, but if logistical considerations are
excluded in advance, the same network will probably suffice. As pressure on
the light, the SPARCstation 10 Model 40 is more than adequate. (Even in the very
bad case, there are only two networks). Processing capacity of SPARCclassic also
usually sufficient. While the total number of very small (25 MB of source code
and object code 18 megabytes; even 20 complete copies of only 660 megabytes),
the recommended disk configuration can include two 535 MB on disk. The
assumption is that CFS can be and one disc, as header files are often read from
the server (they cache clients). When one or two drives one of the
SCSI bus full enough. Volume data is very small and most of them will be read
and sent many clients repeatedly, it is certainly enough to configure memory for
all the data cache : 16 MB of memory under the basic operating system, plus a 25
MB cache for the source code in the configuration of 48-64 MB. Because of this
environment writes a frequent NVSIMM or PrestoServe are significant. The final
version of the station can choose either entry-level SPARCstation 10, or well-configured
stations SPARCclassic. For the common sense to see that the maximum speed of 166
queries per ops / sec at 75% less than that of SPARCclassic (236 ops / sec) to
test LADDIS (remember that the speed of 166 ops / sec assumed that all 20
clients fully active at all times, although the use of real logs show that this
never happens); Maximum load required half of which shows SPARCstation 10 Model
40 to test LADDIS (411 ops / sec). Comparison with the corresponding figure
LADDIS situation intensive attributes, as results LADDIS intensive use of a
mixture of attributes. Table 4.3. The LADDIS different NFS-serverov Sun with
Solaris 2.3. A few (5%) higher speeds achieved with the use of FDDI, if there is
no system for the measurement and application behavior is not well understood,
it can be assessed based on a similar application load, as shown in Tables 4.4-4.6.
These figures give some indication of the measured loads and NFS. This does not
mean that they give a picture of what to expect from the burden of certain tasks.
In particular, note that the data contained in these tables are the maximum
anticipated load of real clients, as these figures reflect only the period of
time when the system is active NFS-zaprosy. As noted above, sect. 3.1.4, the
system almost never fully active at all times. A notable exception to this rule
is computer servers, which in reality are constantly working packet machine. For
example, the work of 486/33, serving 1-2-3 is shown in Table 4.2 and figure.
4.2. Although presented in the table peak load is 80 ops / sec, it is clear from
the figure that the total load is less than 10% of the speed and the average
five minutes, much less load 10 ops / sec. The average for a longer period of
time, pressure is approximately 0.1 PC ops / sec. Most workstations class
SPARCstation2 or SPARCstation ELC are averaging 1 op / sec, and most reasonable
equivalents customers SPARCstation 10 Model 51, Model 512, HP 9000/735 or RS6000/375
1-2 ops / sec. Of course, these figures vary significantly depending on the
individual user and application.
|